PC Monitors

HP Performance Center
for the Windows operating systems

Software Version: 9.50
Monitor Reference
Document Number: FILL IN
Document Release Date: January 2009
Software Release Date: January 2009
2
Legal Notices
Warranty
The only warranties for HP products and services are set forth in the express warranty
statements accompanying such products and services. Nothing herein should be construed as
constituting an additional warranty. HP shall not be liable for technical or editorial errors or
omissions contained herein.
The information contained herein is subject to change without notice.
Restricted Rights Legend
Confidential computer software. Valid license from HP required for possession, use or copying.
Consistent with FAR 12.211 and 12.212, Commercial Computer Software, Computer Software
Documentation, and Technical Data for Commercial Items are licensed to the U.S.
Government under vendor's standard commercial license.
Third-Party Web Sites
HP provides links to external third-party Web sites to help you find supplemental
information. Site content and availability may change without notice. HP makes no
representations or warranties whatsoever as to site content or availability.
Copyright Notices
Copyright 2002 - 2009 Mercury Interactive (Israel) Ltd.
Trademark Notices
Intel is a trademark of Intel Corporation in the U.S. and other countries.
Java is a US trademark of Sun Microsystems, Inc.
Microsoft and Windows are U.S. registered trademarks of Microsoft Corporation.
Oracle is a registered US trademark of Oracle Corporation, Redwood City, California.
UNIX is a registered trademark of The Open Group.
3
Documentation Updates
This guides title page contains the following identifying information:
Software Version number, which indicates the software version.
Document Release Date, which changes each time the document is updated.
Software Release Date, which indicates the release date of this version of the software.
To check for recent updates, or to verify that you are using the most recent edition of a
document, go to:
http://h20230.www2.hp.com/selfsolve/manuals
This site requires that you register for an HP Passport and sign-in. To register for an HP
Passport ID, go to:
http://h20229.www2.hp.com/passport-registration.html
Or click the New users - please register link on the HP Passport login page.
You will also receive updated or new editions if you subscribe to the appropriate product
support service. Contact your HP sales representative for details.
4
Support
You can visit the HP Software Support web site at:
http://www.hp.com/go/hpsoftwaresupport
This web site provides contact information and details about the products, services, and
support that HP Software offers.
HP Software Support Online provides customer self-solve capabilities. It provides a fast and
efficient way to access interactive technical support tools needed to manage your business. As
a valued support customer, you can benefit by using the HP Software Support web site to:
Search for knowledge documents of interest
Submit and track support cases and enhancement requests
Download software patches
Manage support contracts
Look up HP support contacts
Review information about available services
Enter into discussions with other software customers
Research and register for software training
Most of the support areas require that you register as an HP Passport user and sign in. Many
also require a support contract.
To find more information about access levels, go to:
http://h20230.www2.hp.com/new_access_levels.jsp
To register for an HP Passport ID, go to:
5
Table of Contents
Welcome to This Guide.......................................................................13
How This Guide Is Organized .............................................................13
Who Should Read This Guide .............................................................16
Performance Center Documentation Library .....................................16
Additional Online Resources...............................................................18
Documentation Updates .....................................................................19
PART I : I NTRODUCTI ON
Chapter 1: Introducing Performance Center Monitoring..................23
Setting Up the Monitoring Environment ...........................................24
Configuring Performance Center Monitors on the Controller ..........26
Configuring the Remote Machine for SiteScope Monitors.................26
PART I I : RUN-TI ME AND TRANSACTI ON MONITORI NG
Chapter 2: Run-Time Monitoring........................................................39
Run-Time Graphs ................................................................................39
Chapter 3: Transaction Monitoring....................................................43
Transaction Monitor Graphs...............................................................43
Enabling the Transaction Monitor .....................................................44
Adding Transactions to a Script .........................................................45
Table of Contents
6
PART I I I: WEB RESOURCE MONI TORI NG
Chapter 4: Web Resource Monitoring................................................49
About Web Resource Monitoring........................................................49
Hits per Second Graph ........................................................................50
Throughput Graph .............................................................................50
HTTP Responses per Second Graph ...................................................51
Pages Downloaded per Second Graph ...............................................53
Retries per Second Graph ...................................................................54
Connections Graph ............................................................................55
Connections per Second Graph .........................................................55
SSL Connections per Second Graph ...................................................56
PART I V: SYSTEM RESOURCE MONI TORI NG
Chapter 5: Introducing System Resource Monitoring........................59
Chapter 6: Antara FlameThrower Monitoring....................................61
Configuring the Antara FlameThrower Monitor ................................61
Antara FlameThrower Performance Counters ....................................63
Chapter 7: Windows Resource Monitoring ........................................75
Configuring the Windows Resources Monitor ..................................76
Windows Resources Performance Counters........................................79
Chapter 8: SNMP Resource Monitoring .............................................83
Configuring the SNMP Resources Monitor.........................................83
Configuring the SiteScope SNMP by MIB Monitor.............................87
Chapter 9: SiteScope Resource Monitoring........................................91
Before Setting Up the SiteScope Monitor............................................91
Configuring the SiteScope Monitor ....................................................92
Chapter 10: Server Resource Monitoring...........................................97
Setting Up the Server Resources Monitor............................................98
Configuring the Server Resources Monitor.........................................98
Server Resources Performance Counters ...........................................100
Chapter 11: UNIX Resource Monitoring...........................................103
Setting Up the Monitoring Environment .........................................104
Configuring the UNIX Resources Monitor .......................................106
UNIX Resources Performance Counters............................................108
Table of Contents
7
PART V: NETWORK MONI TORI NG
Chapter 12: Introducing Network Monitoring.................................111
Chapter 13: Network Delay Monitoring...........................................113
Network Monitoring from a UNIX Source Machine ........................113
Configuring the Network Delay Time Monitor ................................117
Network Monitoring over a Firewall .................................................121
PART VI : FIREWALL SERVER MONI TORING
Chapter 14: Firewall Server Performance Monitoring .....................125
About the Firewall Server Monitor....................................................125
Configuring the Check Point FireWall-1 Server Monitor .................125
Check Point FireWall-1 Performance Counters ................................127
PART VI I : WEB SERVER MONI TORI NG
Chapter 15: Introducing Web Server Monitoring............................131
About Web Server Resource Monitors...............................................131
Monitoring Using a Proxy Server .....................................................132
Chapter 16: Apache Resource Monitoring .......................................135
Configuring the Apache Monitor ....................................................135
Apache Performance Counters..........................................................137
Chapter 17: Microsoft IIS Monitoring ..............................................139
Configuring the Microsoft IIS Monitor ............................................139
Microsoft IIS Performance Counters .................................................142
Chapter 18: iPlanet/Netscape Monitoring .......................................143
Configuring the iPlanet/Netscape Monitor .....................................144
iPlanet/Netscape Performance Counters...........................................147
Chapter 19: iPlanet (SNMP) Monitoring..........................................149
Configuring the iPlanet (SNMP) Monitor.........................................149
iPlanet (SNMP) Performance Counters .............................................151
PART VI I I : WEB APPLI CATI ON SERVER RESOURCE MONITORI NG
Chapter 20: Introducing Web Application Server
Resource Monitoring ....................................................................159
Chapter 21: Ariba Monitoring ..........................................................161
Configuring the Ariba Monitor.........................................................161
Ariba Performance Counters .............................................................164
Table of Contents
8
Chapter 22: ATG Dynamo Monitoring .............................................167
Configuring the ATG Dynamo Monitor ..........................................167
ATG Dynamo Performance Counters ...............................................170
Chapter 23: BroadVision Monitoring ...............................................173
Configuring the BroadVision Monitor ............................................174
BroadVision Servers and Services ......................................................177
BroadVision Performance Counters..................................................179
Chapter 24: ColdFusion Monitoring.................................................187
Configuring the ColdFusion Monitor ..............................................188
ColdFusion Performance Counters ...................................................190
Chapter 25: iPlanet (NAS) Monitoring.............................................191
Configuring the iPlanet (NAS) Monitor............................................196
iPlanet (NAS) Performance Counters ................................................198
Chapter 26: Microsoft Active Server Pages Monitoring...................209
Configuring the Microsoft Active Server Pages Monitor .................209
MS Active Server Pages Performance Counters.................................212
Chapter 27: Oracle9iAS HTTP Server Monitoring ............................213
Configuring the Oracle9iAS HTTP Server Monitor ..........................213
Oracle9iAS HTTP Server Performance Counters ...............................217
Chapter 28: SilverStream Monitoring...............................................223
Configuring the SilverStream Monitor ............................................223
SilverStream Performance Counters..................................................226
Chapter 29: WebLogic (JMX) Monitoring ........................................227
Configuring the WebLogic (JMX) Monitor.......................................231
BEA WebLogic (JMX) Performance Counters ...................................235
Chapter 30: WebSphere Monitoring................................................237
Configuring the WebSphere Monitor ...............................................238
WebSphere Performance Counters ...................................................243
Chapter 31: WebSphere Application Server Monitoring.................251
Setting up the Monitoring Environment..........................................252
Configuring the WebSphere Application Server Monitor ................254
Chapter 32: WebSphere (EPM) Monitoring.....................................259
Configuring the WebSphere (EPM) Monitor ...................................267
Table of Contents
9
PART I X: DATABASE RESOURCE MONITORI NG
Chapter 33: Introducing Database Resource Monitoring................271
Chapter 34: SQL Server Monitoring .................................................273
Configuring the SQL Server Monitor ...............................................273
SQL Server Performance Counters ....................................................276
Chapter 35: DB2 Database Server Monitoring.................................279
Configuring the DB2 Monitor ..........................................................281
DB2 Performance Counters...............................................................282
Chapter 36: Oracle Database Server Monitoring.............................297
Setting Up the Oracle Monitor Environment ...................................298
Configuring Oracle Database Resource Monitoring .........................302
Oracle Performance Counters ...........................................................306
Custom Queries .................................................................................307
Chapter 37: Sybase Database Server Monitoring ............................309
Configuring the Sybase Monitor.......................................................311
Sybase Performance Counters ...........................................................314
PART X: J 2EE PERFORMANCE MONI TORI NG
Chapter 38: J2EE Performance Monitoring ......................................323
About J2EE Performance Monitoring................................................323
Setting Up the Monitoring Enviroment............................................324
Configuring the J2EE Monitor ..........................................................329
J2EE Performance Counters...............................................................331
Examples of Modifying Application Server Configuration...............332
PART XI : APPLI CATI ON DEPLOYMENT MONI TORI NG
Chapter 39: Application Deployment Solutions ...............................345
Configuring the Citrix MetaFrame Server Monitor ..........................347
Citrix MetaFrame Performance Counters .........................................350
Table of Contents
10
PART XI I : ERP/CRM SERVER RESOURCE MONI TORI NG
Chapter 40: Introducing ERP/CRM Server Resource Monitoring ....359
Chapter 41: SAPGUI Monitoring.......................................................361
Configuring the SAPGUI Monitor.....................................................362
SAPGUI Performance Counters.........................................................365
Chapter 42: SAP Portal Monitoring..................................................367
Configuring the SAP Portal Monitor.................................................367
SAP Portal Performance Counters .....................................................371
Chapter 43: Siebel Web Server Monitoring .....................................373
Configuring the Siebel Web Server Monitor.....................................374
Siebel Web Server Performance Counters .........................................378
Chapter 44: Siebel Server Manager Monitoring ..............................379
Setting up the Monitoring Environment..........................................379
Configuring the Siebel Server Manager Monitor ..............................381
Siebel Server Manager Performance Counters ..................................382
Chapter 45: PeopleSoft (Tuxedo) Monitoring .................................385
Setting Up the PeopleSoft (Tuxedo) Monitor ...................................385
Configuring the PeopleSoft (Tuxedo) Monitor.................................386
PeopleSoft (Tuxedo) Performance Counters .....................................389
PART XI I I: APPLI CATI ON COMPONENT MONITORI NG
Chapter 46: Introducing Application Component Monitoring........393
Chapter 47: Microsoft COM+ Monitoring........................................395
Configuring the Microsoft COM+ Server Monitor ...........................396
Microsoft COM+ Performance Counters ..........................................398
PART XI V: MI DDLEWARE PERFORMANCE MONI TORI NG
Chapter 48: Middleware Performance Monitoring..........................405
Configuring the Tuxedo Monitor .....................................................407
Tuxedo Monitor Measurements........................................................410
Table of Contents
11
PART XV: TROUBLESHOOTING MONI TORS
Chapter 49: Troubleshooting Monitors............................................415
Cannot Create a Monitor ..................................................................415
Cannot Save Measurements ..............................................................417
Port Required for J2EE Monitor Is Already Taken.............................418
J2EE Monitor Initialization Errors.....................................................419
Cannot Save SiteScope Monitors with Many Counters....................419
Index..................................................................................................421
Table of Contents
12
13
Welcome to This Guide
Welcome to the HP Performance Center Monitor Reference.
Performance Center is HPs Web-enabled global load testing tool, which is
specially designed to streamline the testing process and increase the test
efficiency for multiple concurrent load tests across multiple geographic
locations.
This guide describes how to configure the HP Performance Center monitors
and SiteScope monitors, and lists each monitors performance counters.
This chapter includes:
How This Guide Is Organized on page 13
Who Should Read This Guide on page 16
Performance Center Documentation Library on page 16
Additional Online Resources on page 18
Documentation Updates on page 19
How This Guide Is Organized
The HP Performance Center Monitor Reference contains the following sections:
Part I Introduction
Introduces Performance Center monitoring and setting up monitoring
environments.
14
Part II Run-Time and Transaction Monitoring
The Run-Time and Transaction monitors display information related to the
Vusers and transactions during the load test run.
Part III Web Resource Monitoring
The Web Resource monitor measures statistics at the Web servers during
load test runs. It provides information about the number of Web
connections, throughput volume, HTTP responses, server retries, and
downloaded pages during the load test.
Part IV System Resource Monitoring
The System Resource monitors monitor a machines system resource usage
during a load test run. They guage the Antara FlameThrower, Windows,
Unix, SiteScope server resources used during a load test.
Part V Network Monitoring
The Network monitor displays information about the network delays on
your system.
Part VI Firewall Server Monitoring
The Firewall monitor measures statistics at the firewall servers during the
load test.
Part VII Web Server Monitoring
The Web Server resource monitors measure statistics on the Apache,
Microsoft IIS, iPlanet (SNMP) and iPlanet/Netscape Web servers during the
load test.
Part VIII Web Application Server Resource Monitoring
The Web Application Server Resource monitors measure statistics on the
Ariba, ATG Dynamo, BroadVision, ColdFusion, iPlanet (NAS), MS Active
Server Pages, Oracle9iAS, SilverStream, WebLogic (JMX), and WebSphere
application servers during the load test.
15
Part IX Database Resource Monitoring
The Database Server Resource monitors measure statistics related to the
SQL, DB2, Sybase, and Oracle servers.
Part X J2EE Performance Monitoring
The Java Performance monitor measure statistics of Java 2 Platform,
Enterprise Edition (J2EE) objects using the J2EE server machine.
The J2EE/.NET Diagnostics monitors provide information to trace, time, and
troubleshoot individual transactions through J2EE/.NET Web, application,
and database servers. For information about the J2EE/.NET Diagnostics
monitors, see the HP Diagnostics documentation.
Part XI Application Deployment Monitoring
The Application Deployment Solutions monitor measures statistics of the
Citrix MetaFrame XP and 1.8 servers during a load test run.
Part XII ERP/CRM Server Resource Monitoring
The ERP/CRM Server Resource monitors measure statistics of the SAP,
SAPGUI, SAP Portal, Siebel Server Manager, Siebel Web Server, and
PeopleSoft (Tuxedo) server during a load test run.
Part XIII Application Component Monitoring
The Application Component monitors measure statistics of the Microsoft
COM+ server during a load test run.
Part XIV Middleware Performance Monitoring
The Middleware Performance monitor measures statistics of the Tuxedo
server during a load test run.
Part XV Troubleshooting Monitors
This section provides troubleshooting for problems related to working with
the Performance Center monitors.
16
Who Should Read This Guide
This guide is for the following users of Performance Center:
Performance Engineers
Project Managers
This document assumes that you are moderately knowledgeable about
enterprise application development and highly skilled in enterprise system
and database administration.
Performance Center Documentation Library
Performance Center includes a complete set of documentation describing
how to use the product. A comprehensive Documentation Library is
available from the Help menu in the User and Administration Sites. PDFs
can be read and printed using Adobe Reader, which can be downloaded
from the Adobe Web site (http://www.adobe.com). Printed documentation is
also available on demand.
Accessing the Documentation
You can access Performance Center documentation as follows:
In the User and Administration Sites, click the Help link in the top right
corner of the window. In the bookmarks on the left, click Home to open
the Performance Center Documentation Library front page which
provides quick links to the online Performance Center Documentation
Library.
Printable guides can be accessed from Start > Program
Files > Performance Center > Documentation.
Getting Started Documentation
Readme. Provides last-minute news and information about Performance
Center.
17
HP Performance Center System Configuration and Installation Guide.
Explains how to install and configure the Performance Center
components. Available in PDF format only. Accessible from the HP
Installation DVD.
HP Performance Center Quick Start. A self-paced guide showing you how
to use Performance Center to create, run, and monitor load tests.
Available in PDF format only.
Accessible from the General tab that is displayed when you open the User
Site.
Performance Center User Guides
HP Performance Center User Guide. Describes how to use Performance
Center to create, schedule, run, and monitor load tests.
HP Performance Center Monitor Reference. Describes how to set up the
server monitor environment and configure Performance Center monitors
for monitoring data generated during a load test run.
Performance Center Administration
HP Performance Center Administrator Guide. Describes how to use
Performance Center Administration Site for overall resource
management, site management, system configuration, technical
supervision, and user privileges in Performance Center projects.
Troubleshooting
HP Performance Center Troubleshooting Guide. Provides information
for troubleshooting problems while working with Performance Center
Available in PDF fomat only.
Standalone Applications
The following documentation is available only on host machines, and if
standalone components are installed:
HP Analysis User Guide. Describes how to generate graphs and reports to
analyze your load tests after they have run.
18
HP Virtual User Generator User Guide. Describes how to create Vuser
scripts through recording and programming.
Additional Online Resources
Troubleshooting and Knowledge Base accesses the Troubleshooting page on
the HP Software Support Web site where you can search the Self-solve
knowledge base. The URL for this Web site is
http://h20230.www2.hp.com/troubleshooting.jsp.
HP Software Support accesses the HP Software Support Web site. This site
enables you to browse the Self-solve knowledge base. You can also post to
and search user discussion forums, submit support requests, download
patches and updated documentation, and more. The URL for this Web site is
www.hp.com/go/hpsoftwaresupport.
Most of the support areas require that you register as an HP Passport user
and sign in. Many also require a support contract.
To find more information about access levels, go to:
http://h20230.www2.hp.com/new_access_levels.jsp
To register for an HP Passport user ID, go to:
HP Software Web site accesses the HP Software Web site. This site provides
you with the most up-to-date information on HP Software products. This
includes new software releases, seminars and trade shows, customer support,
and more. The URL for this Web site is www.hp.com/go/software.

19
Documentation Updates
HP Software is continually updating its product documentation with new
information.
To check for recent updates, or to verify that you are using the most recent
edition of a document, go to the HP Software Product Manuals Web site
(http://h20230.www2.hp.com/selfsolve/manuals).
20
Part I
Introduction
22
23
1
Introducing Performance Center
Monitoring
You can monitor a load test run using the Performance Center online
monitors.
Setting Up the Monitoring Environment on page 24
Configuring Performance Center Monitors on the Controller on page 26
Configuring the Remote Machine for SiteScope Monitors on page 26
Chapter 1 Introducing Performance Center Monitoring
24
Setting Up the Monitoring Environment
Before monitoring a load test run, you need to set up and configure the
Performance Center monitoring components. Each monitor has different
configuration requirements that are explained in the specific monitoring
chapters. The diagram below illustrates the Performance Center monitoring
process.
Before monitoring a server, perform the following steps:
configure the monitoring environment on the server machine (if
necessary)
configure the monitor on the Controller machine
25
Configuring the Monitoring Environment on the Server
Machine
To use the following monitors, you must first install or configure
monitoring components on the server machine:
Citrix
DB2
iPlanet (NAS)
J2EE
Microsoft .NET CLR
Microsoft COM+
Network Delay
Oracle
PeopleSoft (Tuxedo)
SAPGUI
SAP Portal
Siebel Server Manager
Siebel Web Server
SiteScope
Sybase
Tuxedo
UNIX
WebLogic (JMX)
WebSphere
WebSphere Application
Server
WebSphere (EPM)
26
Configuring Performance Center Monitors on the
Controller
To obtain performance data for a monitor, you need to configure the
monitor (from the Controller), and indicate which statistics and
measurements you want to monitor. You select these counters using the
monitors Add Measurements page.
To set up a monitor, you need to perform the following:
1 Select monitor.
Select the monitor that you want to measure from the Add Monitor page.
2 Select server.
Add a monitored machine to the Controller by selecting the server whose
monitors you want to configure.
For SiteScope monitors, configure the remote machine.
3 Select measurements.
Configure the monitor by selecting the measurements that you want to
monitor.
For more information on setting up the monitoring environment and
configuring a monitor, see the specific monitoring chapter.
Configuring the Remote Machine for SiteScope Monitors
You configure the remote machine according to the platform your machine
is running on. This section describes configuring remote machines running
on NT and UNIX.
This section includes:
Configuring the NT Remote Machine on page 27
Configuring the UNIX Remote Machine on page 31
27
Configuring the NT Remote Machine
For machines on a NT platform, when you add measurements to monitor
for the first time, you need to configure the remote machine properties. You
configure the remote machine properties in the Configuring NT Remote
Machine page.
To configure the NT remote machine:
1 After you type monitor information for the monitor and click Next, the
Configuring NT Remote Machine dialog box opens.
28
2 Enter the remote machines configuration information.
NT Server Address. The IP address or UNC style name of the NT server
you wish to monitor. An IP host name will also work provided that the
SiteScope server has a way to resolve this common name into an IP
address (for example, by the use of a hosts file, DNS, or WINS/DNS
integration).
Connection Method. SiteScope can use one of two connection types for
monitoring remote NT server resources. These are:
NetBIOS. The default server-to-server communication protocol for
Windows NT and 2000 networks.
SSH. Secure Shell, a more secure communication protocol that can be
installed on Windows NT/2000 based networks. This connection
method normally requires installing SSH libraries on each server to
which you want to connect.
Login. The login for the remote server. If the server is within the same
domain as the SiteScope machine, include the domain name in front of
the user login name. For example, domainname\user. If you are using a
local machine login account for machines within or outside the domain,
include the machine name in front of the user login name. For example,
machinename\user.
Password. The password for the remote server or the passphrase for the
SSH key file.
When using SSH authentication with public/private key based
authentication type the passphrase for the identity file here.
Title. (Optional) A name by which the remote machine should be
known. This name will appear in the drop-down list.
Trace. Check this box to have trace messages to and from the subject
server recorded to the SiteScope RunMonitor.log file.
SSH Connection Method. The method to use for this connection. The
currently supported methods are:
Internal Java Libraries. Connect using the Java SSH client integrated
with SiteScope.
29
Plink. Connect using an external SSH client. On NT, SiteScope ships
with Plink. On UNIX or Linux SiteScope will use an installed client
such as OpenSSH.
Disable Connection Caching. Check this option to turn off connection
caching for this remote. By default SiteScope caches open connections.
Connection Limit. Controls the number of open connections that
SiteScope will allow for this remote. If you have a large number of
monitors configured to use this connection then set this number high
enough to relieve the potential bottleneck.
Note: This setting does not effect the running of tests for a remote, tests
will always create a new connection.
SSH Authentication Method. The authentication method to use for SSH
connections. The currently supported methods are:
Password. Authenticate using a password.
Key File. Authenticate using public/private key authentication. When
this option is selected SiteScope uses the private key in the file
SiteScope/groups/identity to authenticate. The corresponding public
key must be listed in the authorized_keys file on the remote host.
Key File for SSH connections. Select the file that contains the private key
for this connection. The default key file is SiteScope\groups\identity.
This setting only applies when the authentication method is Key File.
SSH Version 2 Only. Check this option to force SiteScope to use SSH
protocol version 2 only. This option only applies when using the
integrated Java Client in SiteScope.
Custom Commandline. Type a custom commandline for a remote using
the External Client. This option can be used when needing to pass
specific options to the external client being executed. Valid substitution
variable are:
$root$. This will be translated to the SiteScope directory.
$user$. This will be translated to the user name entered into the
remote.
30
$password$. This will be translated to the password entered into the
remote.
$host$. This will be translated to the host name entered into the
remote.
SSH Port Number. Type the port that the remote SSH server is listening
on. By default, the port number is 22.
3 Click Next, and continue to add measurements in the Choose
Measurements page.
31
Configuring the UNIX Remote Machine
For machines on a UNIX platform, when you add measurements to monitor
for the first time, the Configuring Unix Remote Machine page opens.
To configure the UNIX remote machine:
1 After you enter monitor information for the monitor and click Next, the
Configuring UNIX Remote Machine dialog box opens.
2 Enter the Unix Remote Machine information:
Server Address. Displays the IP address or host name of the server that
you entered in the Add Machine dialog box.
32
OS. Select the operating system running on the remote server. The
following versions of UNIX are supported:
Connection Method. Select the method for connecting to the server. The
supported methods are:
Telnet. Log in to the remote server using Telnet.
SSH. Log in to the remote server using the SSH protocol. This may
require additional software and setup depending on the version of
UNIX you are working with.
Rlogin. Log in to the remote server using the Rlogin protocol.
HTTP. Connect to an HTTP server on the remote server and run the
command via a CGI. For this method, the Login and Password are
optional and are used for authorizing the log on to the remote
machine if required.
Login. The login for the remote server.
Password. Type the password for the remote server.
Title. Type a name by which the remote machine should be known. This
name will appear in the list in monitors that can connect to this server.
Prompt. Type the prompt to be displayed when the system is ready to
handle a command. The default is #.
Login Prompt. Type the prompt to be displayed when the system is
waiting for the login to be entered. The default is "ogin:".
Password Prompt. Type the prompt to be displayed when the system is
waiting for the password to be entered. The default is "assword:".
AIX OPENSERVER
FreeBSD SCO
HP/UX SGI Irix
HP/UX 64-bit Sun Solaris
Linux Tru64 5.x
MacOSX Tru64 Pre 4.x (Digital)
33
Secondary Prompt. Type the prompt to be displayed if the telnet
connection to the remote server causes the remote server to prompt for
more information about the connection. Separate multiple prompt
strings by commas (,). For example, for Telnet connections to some
remote servers, the remote server may ask what terminal type should be
emulated for the connection. In this case you might need to type
Terminal type? as the secondary prompt. The response to the secondary
prompt is entered in the Secondary Response field below.
Secondary Response. Type the responses to secondary prompts required
to establish connections with this remote server. Separate multiple
responses with commas (,).
Initialize Shell Environment. Type any shell commands to be executed at
the beginning of the session. Separate multiple commands with a
semicolon (;). This option allows you to specify shell commands to be
executed on the remote machine directly after a Telnet or SSH session has
been initiated. These commands can be used to customize the shell for
each SiteScope remote.
Examples:
The remote shell may not have the correct path set for SiteScope
scripts to run. The following command will add the directory
/usr/local/bin into the PATH of the current shell on the remote
machine:
The remote shell may not be initializing the pseudo terminal correctly.
Type the following command to increase the terminal width to 1024
characters:
Note: Commands after a shell invocation will not be executed.
export PATH=$PATH:/usr/local/sbin
stty cols 1024;${SHELL}
34
There have been cases where the remote Telnet Server does not echo
back the command line properly. This may cause strange behavior for
monitors that rely on this behavior.
Type the following command to force the remote terminal to echo:
Certain UNIX shells have been known to behave erratically with
SiteScope. This includes bash, ksh, and csh. Type the following
command to change the shell to sh for the SiteScope connection:
Trace. Select this option to trace messages to and from the remote server
in the RunMonitor.log file.
SSH Connection Method. Select the method to use to connect to the
remote server.
Internal Java Libraries. Connect using the Java SSH client integrated
with SiteScope.
Plink. Connect using an external SSH client. On Windows NT,
SiteScope ships with Plink.
Disable Connection Caching. Select this to disable SSH connection
caching.
Connection Limit. Type the maximum number of connections for this
remote machine.
SSH Authentication Method. Select the method to use to authenticate to
the remote server (for SSH connections only).
Password. Authenticate using a password.
Keyfile. Authenticate using public/private key authentication. When
this option is selected, SiteScope uses the private key in the file
SiteScope/groups/identity to authenticate. The corresponding public
key must be listed in the authorized_keys file on the remote host.
Key File for SSH connections. Type the Key File for SSH connections.
stty echo
/bin/sh
35
SSH Version 2 Only. Select this option to force SSH to only use SSH
protocol version 2. This option is only supported when using the internal
Java libraries connection method.
Custom Commandline. Type the command for execution of the external
SSH client. For substitutions with options listed above, use $host$,
$user$, and $password$ respectively. This setting is supported only for
connections using an external process.
SSH Port Number. Type the port on which the SSH service is running.
After defining the server for SiteScope, you can test the settings by clicking
on the test link. SiteScope attempts to display the working directory of the
remote machine (the "pwd" command on UNIX, or "cd" on Windows NT), as
a test to check that the remote machine can be accessed and can run
commands properly.
3 Click Next, and continue to add measurements in the Choose
Measurements page.
36
Part II
Run-Time and Transaction Monitoring
38
39
2
Run-Time Monitoring
While running a load test, you can use Performance Centers Run-Time
monitor to view graphs of the run-time status.
Run-Time Graphs on page 39
Run-Time Graphs
The Run-Time monitor provides information about the status of the Vusers
participating in the load test, and the number and types of errors that the
Vusers generate. In addition, the Run-Time monitor provides the User-
Defined Data Points graph, which displays the real time values for user-
defined points in a Vuser script.
The Run-Time monitor provides information for the following graphs:
Running Vusers Graph
Error Statistics Graph
Vusers with Error Statistics Graph
User-Defined Data Points Graph
Chapter 2 Run-Time Monitoring
40
Running Vusers Graph
The monitors Running Vusers graph provides information about the status
of the Vusers running in the current load test on all load generator
machines. The graph shows the number of running Vusers, while the
information in the legend indicates the number of Vusers in each state.
The Status field of each Vuser displays the current status of the Vuser. The
following table describes each Vuser status:
Error Statistics Graph
The monitors Error Statistics graph provides details about the number of
errors that accrue during each second of the load test run. The errors are
grouped by error sourcefor example, the location in the script or the load
generator name.
Vusers with Error Statistics Graph
The Vusers with Error Statistics graph provides details about the number of
Vusers that generate errors during load test execution. The errors are
grouped by error source.
Status Description
RUNNING The total number of Vusers currently running on all load
generators.
READY The number of Vusers that completed the initialization
section of the script and are ready to run.
FINISHED The number of Vusers that have finished running. This
includes both Vusers that passed and failed.
ERROR The number of Vusers whose execution generated an error.
Check the Status field in the Vuser view or the Output
dialog box for a complete explanation of the error.
41
User-Defined Data Points Graph
The User-Defined Data Points graph displays the real-time values of user-
defined data points. You define a data point in your Vuser script by inserting
an lr_user_data_point function at the appropriate place (user_data_point
for GUI Vusers and lr.user_data_point for Java Vusers).
For Vuser protocols that support the graphical script representations such as
Web and Oracle NCA, you insert a data point as a User Defined step. Data
point information is gathered each time the script executes the function or
step. For more information about data points, see the HP LoadRunner Online
Function Reference.
By default, Performance Center displays all of the data points in a single
graph. The legend provides information about each data point. If desired,
you can hide specific data points using the legend below the graphs.
You can also view data points offline, after the completion of the load test.
For more information, see the HP LoadRunner Analysis User Guide.
Action1()
{
lr_think_time(1);
lr_user_data_point ("data_point_1",1);
lr_user_data_point ("data_point_2",2);
return 0;
}
42
43
3
Transaction Monitoring
While running a load test, you can use Performance Centers Transaction
monitors to view graphs of transaction performance.
Transaction Monitor Graphs on page 43
Enabling the Transaction Monitor on page 44
Adding Transactions to a Script on page 45
Transaction Monitor Graphs
The Transaction monitor displays the transaction rate and response time
during load test execution. For more information about transactions, see
Adding Transactions to a Script on page 45.
The Transaction monitor provides the following graphs:
Transaction Response Time Graph
Transactions per Second (Passed) Graph
Transactions per Second (Failed, Stopped) Graph
Total Transactions per Second (Passed) Graph
Transaction Response Time Graph
The Transaction Response Time graph shows the average response time of
transactions in seconds (y-axis) as a function of the elapsed time in the load
test (x-axis).
Chapter 3 Transaction Monitoring
44
Transactions per Second (Passed) Graph
The Transactions per Second (Passed) graph shows the number of successful
transactions performed per second (y-axis) as a function of the elapsed time
in the load test (x-axis).
Transactions per Second (Failed, Stopped) Graph
The Transactions per Second (Failed, Stopped) graph shows the number of
failed and stopped transactions per second (y-axis) as a function of the
elapsed time in the load test (x-axis).
Total Transactions per Second (Passed) Graph
The Total Transactions per Second (Passed) graph shows the total number
of completed, successful transactions per second (y-axis) as a function of the
elapsed time in the load test (x-axis).
Enabling the Transaction Monitor
The Transaction monitor is enabled by defaultit automatically begins
monitoring Vuser transactions at the start of a load test. To conserve
resources, you can disable the Transaction monitor.
To enable the transaction monitoring:
1 On the Projects menu, click Options.
45
2 In the left pane of the Options page, click Monitors.
3 Under Transaction Data, select Enable Transaction Monitor.
4 Click Save.
Adding Transactions to a Script
If there are no transactions defined in your Vuser script, no data will be
displayed in the online graphs. To add transactions to an existing script, edit
it using the appropriate tool. The following table shows the script
generation tools for each script type:

Script type Editing tool
GUI Windows HP WinRunner
non-GUI Windows HP Virtual User Generator (VuGen)
46
To add a transaction to a script:
1 On the Projects menu, click Vuser Scripts to open the Vuser Scripts page.
2 Click the Download Script arrow. Performance Center prepares the script for
download and opens the Ready for Download dialog box.
3 Click the Download button. The File Download dialog box opens.
4 Click OK. The Save As dialog opens.
5 Choose a file name and the location into which you want the file
downloaded.
6 Click Save to download the file.
7 Using Vugen open and insert Start and End Transaction functions or
markers throughout your script.
Note: Your version of VuGen may be enabled to upload scripts but not be
enabled to download scripts. If you have trouble downloading scripts you
may have to run the update for your version of VuGen. For more
information, see the section about preparing VuGen to work with
Performance Center in the HP Performance Center User Guide.
Part III
Web Resource Monitoring
48
49
4
Web Resource Monitoring
You can obtain information about the performance of your Web server
using Performance Centers Web Resource monitor.
About Web Resource Monitoring on page 49
Hits per Second Graph on page 50
Throughput Graph on page 50
HTTP Responses per Second Graph on page 51
Pages Downloaded per Second Graph on page 53
Retries per Second Graph on page 54
Connections Graph on page 55
Connections per Second Graph on page 55
SSL Connections per Second Graph on page 56
About Web Resource Monitoring
The Web Resource monitor enables you to analyze the throughput on the
Web server, the number of hits per second that occurred during the load
test, the number of HTTP responses per second, the HTTP status codes
(which indicate the status of HTTP requests, for example, "the request was
successful," "the page was not found") returned from the Web server, the
number of downloaded pages per second, the number of server retries per
second, the number of open TCP/IP connections, the number of new TCP/IP
connections per second, and the number of SSL Connections per second.
Chapter 4 Web Resource Monitoring
50
Hits per Second Graph
The Hits Per Second graph shows the number of hits (HTTP requests) to the
Web server (y-axis) as a function of the elapsed time in the load test (x-axis).
This graph can display the whole step, or the last 60, 180, 600, or 3600
seconds. You can compare this graph to the Transaction Response Time
graph to see how the number of hits affects transaction performance.
Throughput Graph
The Throughput graph shows the amount of throughput on the Web server
(y-axis) during each second of the load test run (x-axis). Throughput is
measured in kilobytes and represents the amount of data that the Vusers
received from the server at any given second. You can compare this graph to
the Transaction Response Time graph to see how the throughput affects
transaction performance.
In the following example, the Transaction Response time graph is compared
with the Throughput graph. It is apparent from the graph that as the
throughput decreases, the transaction response time also decreases. The
peak throughput occurred at approximately 1 minute into the step. The
highest response time also occurred at this time.
51
HTTP Responses per Second Graph
The HTTP Responses per Second graph shows the number of HTTP status
codeswhich indicate the status of HTTP requests, for example, "the request
was successful," "the page was not found"(y-axis) returned from the Web
server during each second of the load test run (x-axis), grouped by status
code. You can group the results shown in this graph by script (using the
"Group By" function) to locate scripts which generated error codes.
The following table displays a list of HTTP status codes:
Code Description
200 OK
201 Created
202 Accepted
203 Non-Authoritative Information
204 No Content
205 Reset Content
206 Partial Content
300 Multiple Choices
301 Moved Permanently
302 Found
303 See Other
304 Not Modified
305 Use Proxy
307 Temporary Redirect
400 Bad Request
401 Unauthorized
402 Payment Required
403 Forbidden
52
For more information on the above status codes and their descriptions, see
http://www.w3.org/Protocols/rfc2616/rfc2616-sec10.html#sec10.
404 Not Found
405 Method Not Allowed
406 Not Acceptable
407 Proxy Authentication Required
408 Request Timeout
409 Conflict
410 Gone
411 Length Required
412 Precondition Failed
413 Request Entity Too Large
414 Request - URI Too Large
415 Unsupported Media Type
416 Requested range not satisfiable
417 Expectation Failed
500 Internal Server Error
501 Not Implemented
502 Bad Gateway
503 Service Unavailable
504 Gateway Timeout
505 HTTP Version not supported
Code Description
53
Pages Downloaded per Second Graph
The Pages Downloaded per Second graph shows the number of Web pages
(y-axis) downloaded from the server during each second of the load test run
(x-axis). This graph helps you evaluate the amount of load Vusers generate,
in terms of the number of pages downloaded.
Note: To view the Pages Downloaded per Second graph, you must select
Pages per second (HTML Mode only) from the scripts run-time settings
Preferences tab before running your load test.
Like throughput, downloaded pages per second is a representation of the
amount of data that the Vusers received from the server at any given second.
The Throughput graph takes into account each resource and its size (for
example, the size of each .gif file, the size of each Web page).
The Pages Downloaded per Second graph takes into account simply the
number of pages.
54
In the following example, the Throughput graph is compared with the Pages
Downloaded per Second graph. It is apparent from the graph that
throughput is not proportional to the number of pages downloaded per
second. For example, between 15 and 16 seconds into the load test run, the
throughput decreased while the number of pages downloaded per second
increased.
Retries per Second Graph
The Retries Per Second graph shows the number of attempted Web server
connections (y-axis) as a function of the elapsed time in the load test (x-
axis). A server connection is retried when the initial connection was
unauthorized, when proxy authentication is required, when the initial
connection was closed by the server, when the initial connection to the
server could not be made, or when the server was initially unable to resolve
the load generators IP address.
55
Connections Graph
The Connections graph shows the number of open TCP/IP connections (y-
axis) at each point in time of the load test (x-axis). Note that one HTML
page may cause the browser to open several connections, when links on the
page go to different Web addresses. Two connections are opened for each
Web server.
This graph is useful in indicating when additional connections are needed.
For example, if the number of connections reaches a plateau, and the
transaction response time increases sharply, adding connections would
probably cause a dramatic improvement in performance (reduction in the
transaction response time).
Connections per Second Graph
The Connections Per Second graph shows the number of new TCP/IP
connections (y-axis) opened each second of the load test (x-axis). This
number should be a small fraction of the number of hits per second, because
new TCP/IP connections are very expensive in terms of server, router and
network resource consumption. Ideally, many HTTP requests should use the
same connection, instead of opening a new connection for each request.
56
SSL Connections per Second Graph
The SSL Connections per Second graph shows the number of new and
reused SSL Connections (y-axis) opened in each second of the load test (x-
axis). An SSL connection is opened by the browser after a TCP/IP connection
has been opened to a secure server.
Because creating a new SSL connection entails heavy resource consumption,
you should try to open as few new SSL connections as possible; once youve
established an SSL connection, you should reuse it. There should be no
more than one new SSL connection per Vuser. If you set your run-time
settings to simulate a new Vuser at each iteration (through the Browser
Emulation tab in the Run-Time Settings menu), you should have no more
than one new SSL connection per Vuser per iteration. Ideally, you should
have very few new TCP/IP and SSL connections each second.
Part IV
System Resource Monitoring
58
59
5
Introducing System Resource Monitoring
You can monitor a machines system resource usage during a load test run
using Performance Centers System Resource monitors.
A primary factor in a transactions response time is its system resource usage.
Using the Performance Center resource monitors, you can monitor Antara
FlameThrower, Windows, SNMP, SiteScope server resources, and UNIX
resources on a machine during a load test run, and determine why a
bottleneck occurred on a particular machine.
The Antara FlameThrower monitor can measure the following
performance counters: Layer, TCP, HTTP, SSL/HTTPS, Sticky SLB, FTP,
SMTP, POP3, DNS, and Attacks.
The Windows measurements correspond to the built-in counters
available from the Windows Performance monitor.
The SNMP monitor is available for monitoring machines using the
Simple Network Management Protocol (SNMP). SNMP monitoring is
platform independent.
The SiteScope monitor can measure server, network, and processor
performance counters. SiteScope can also be used to monitor SAP CCMS
(Computer Center Management System) and SNMP server resources. For
detailed information on all the performance counters that SiteScope can
monitor, see the relevant SiteScope documentation.
The Server Resources monitor shows the CPU, disk space, memory, or
application resources used on remote Windows and UNIX servers during
a load test.
Chapter 5 Introducing System Resource Monitoring
60
The UNIX measurements include those available by the rstatd daemon:
average load, collision rate, context switch rate, CPU utilization,
incoming packets error rate, incoming packets rate, interrupt rate,
outgoing packets error rate, outgoing packets rate, page-in rate, page-out
rate, paging rate, swap-in rate, swap-out rate, system mode CPU
utilization, and user mode CPU utilization.
The System Resource monitors are automatically enabled when you execute
a load test. However, you must specify the machine you want to monitor
and which resources to monitor for each machine. You can also add or
remove machines and resources during the load test run.
61
6
Antara FlameThrower Monitoring
The Antara FlameThrower monitor displays statistics about the resource
usage on the Antara FlameThrower server during the load test run.
The Antara FlameThrower monitor can measure the following performance
counters: Layer, TCP, HTTP, SSL/HTTPS, Sticky SLB, FTP, SMTP, POP3, DNS,
and Attacks.
Configuring the Antara FlameThrower Monitor on page 61
Antara FlameThrower Performance Counters on page 63
Configuring the Antara FlameThrower Monitor
You select the Antara FlameThrower resources you want to monitor using
the Antara FlameThrower Choose Measurements page.
To configure the Antara FlameThrower monitor:
1 Open the Add Monitor window.
Before running a load test: In the Load Test configuration pages
Monitors tab, select the type of monitor profile to create:
To create a monitor profile for the current load test, click Add Local
Profile.
To create a monitor profile that can be used for other load tests, click
Add Profile, type a name and description of the profile and click OK.
For details, see the section about configuring monitors in the HP
Performance Center User Guide.
Chapter 6 Antara FlameThrower Monitoring
62
During a load test run: On the Load Test run page:
Click the Design button. The Load Test Design dialog box opens.
Click Monitors.
For details, see the section about creating monitor profiles while running
a load test in the HP Performance Center User Guide.
2 Under System Resource Graphs, select Antara FlameThrower, then click
Next. The Choose Server page opens.
3 Type the server name or IP address of the machine you want to monitor
according to the following format:
For example, merc1:12135
4 Type the user login name and password.
5 Click Next. The Choose Measurements page opens, displaying the available
measurements and server properties.
6 Select the measurements you want to monitor.
For a list of the available performance counters, see Antara FlameThrower
Performance Counters on page 63.
7 Click Save to add the selected resource measurements to the monitor
profile.
<server name>:<port number>
63
Antara FlameThrower Performance Counters
The following tables describe the counters that can be monitored:
Layer Performance Counters
TCP Performance Counters
Measurement Description
TxBytes The total number of Layer 2 data bytes transmitted.
TxByteRate(/sec) The number of Layer 2 data bytes transmitted per second.
TxFrames The total number of packets transmitted.
TxFrameRate(/sec) The number of packets transmitted per second.
RxBytes The total number of Layer 2 data bytes received.
RxByteRate(/sec) The number of Layer 2 data bytes received per second.
RxFrames The total number of packets received.
RxFrameRate(/sec) The number of packets received per second.
ActiveTCPConns Total number of currently active TCP connections.
SuccTCPConns Total number of SYN ACK packets received.
SuccTCPConn
Rate(/sec)
Number of SYN ACK packets received per second.
TCPConnLatency(mil
isec)
Interval between transmitting a SYN packet and receiving
a SYN ACK reply packet in milliseconds.
MinTCPConn
Latency(msec)
Minimum TCPConnectionLatency in milliseconds.
MaxTCPConn
Latency(msec)
Maximum TCPConnectionLatency in milliseconds.
TCPSndConnClose Total number of FIN or FIN ACK packets transmitted
(Client).
64
HTTP Performance Counters
TCPRcvConnClose Total number of FIN or FIN ACK packets received
(Client).
TCPSndResets Total number of RST packets transmitted.
TCPRcvResets Total number of RST packets received.
SYNSent Total number of SYN packets transmitted.
SYNSentRate(/sec) Number of SYN packets transmitted per second.
SYNAckSent Total number of SYN ACK packets transmitted.
SYNAckRate(/sec) Number of SYN ACK packets transmitted per second.
HTTPRequests Total number of HTTP Request command packets
transmitted.
HTTPRequestRate
(/sec)
Number of HTTP Request packets transmitted per second.
AvgHTTPData
Latency(msec)
The average HTTP Data Latency over the past second in
milliseconds.
HTTPData
Latency(msec)
Interval between transmitting a Request packet and
receiving a response in milliseconds.
DataThroughput
(bytes/sec)
The number of data bytes received from the HTTP server
per second.
MinHTTPData
Latency(msec)
Minimum HTTP Data Latency in milliseconds.
MaxHTTPData
Latency(msec)
Maximum HTTP Data Latency in milliseconds.
MinData
Throughput
(bytes/sec)
Minimum HTTP Data Throughput in seconds.
65
SSL/HTTPS Performance Counters
MaxData
Throughput
(bytes/sec)
Maximum HTTP Data Throughput in seconds.
SuccHTTPRequests Total number of successful HTTP Request Replies
(200 OK) received.
SuccHTTPRequest
Rate(/sec)
Number of successful HTTP Request Replies (200 OK)
received per second.
UnSuccHTTP
Requests
Number of unsuccessful HTTP Requests.
SSLConnections Number of Client Hello messages sent by the Client.
SSLConnection
Rate(/sec)
Number of Client Hello messages sent per second.
SuccSSL
Connections
Number of successful SSL Connections. A successful
connection is one in which the Client receives the
Servers finished handshake message without any errors.
SuccSSLConnectionR
ate(/sec)
Number of successful SSL connections established per
second.
SSLAlertErrors Number of SSL alert messages received by the client (e.g.
bad_record_mac, decryption_failed, handshake_failure,
and so on).
SuccSSLResumed
Sessions
Number of SSL Sessions that were successfully resumed.
FailedSSLResumed
Sessions
Number of SSL Sessions that were unable to be resumed.
66
Sticky SLB Performance Counters
FTP Performance Counters
Cookie
AuthenticationFail
The number of cookies that were not authenticated by
the Server.
SuccCookie
Authentication
The number of cookies authenticated by the server.
SSLClientHellos The number of Client Hello packets sent to the server.
SSLServerHellos The number of Server Hello packets sent to back to the
client.
SSLSessionsFailed The number of Session IDs that were not authenticated
by the server.
SSLSessions
Resumed
The number of Session IDs authenticated by the server.
succSSLClientHellos The number of Client Hello replies received by the client
or packets received by the server.
succSSLServerHellos The number of Server Hellos received by the client.
TPUsers Total number of FTP User command packets transmitted.
FTPUserRate(/sec) Number of FTP User command packets transmitted per
second.
FTPUserLatency
(msec)
Interval between transmitting an FTP User command
packet and receiving a response in milliseconds.
MinFTPUserLatency(
msec)
Minimum FTP Users Latency in milliseconds.
MaxFTPUserLatency(
msec)
Maximum FTP Users Latency in milliseconds.
SuccFTPUsers Total number of successful FTP User command replies
received.
67
SuccFTPUserRate
(/sec)
Number of successful User command replies received per
second.
FTPPasses Total number of FTP PASS packets transmitted.
FTPPassRate(/sec) Number of FTP PASS packets transmitted per second.
FTPPassLatency
(msec)
Interval between transmitting a FTP PASS packet and
MinFTPPassLatency
(msec)
Minimum FTP Pass Latency in milliseconds.
MaxFTPPassLatency(
msec)
Maximum FTP Pass Latency in milliseconds.
SuccFTPPasses Total number of successful FTP PASS replies received.
SuccFTPPassRate
(/sec)
Number of successful FTP PASS replies received per
second.
FTPControl
Connections
Total number of SYN packets transmitted by the FTP
client.
FTPControl
ConnectionRate
(/sec)
Number of SYN packets transmitted by the FTP client per
second.
SuccFTPControl
Connections
Total number of SYN ACK packets received by the FTP
client.
SuccFTPControl
ConnectionRate
(/sec)
Number of SYN ACK packets received by the FTP Client
per second.
FTPData
Connections
Number of SYN ACK packets received by the FTP client
per second.
FTPDataConnectionR
ate(/sec)
Number of SYN ACK packets transmitted by the FTP
Client or received by the FTP Server per second.
SuccFTPData
Connections
Total number of SYN ACK packets transmitted by the FTP
Client or received by the FTP Server.
68
SMTP Performance Counters
SuccFTPData
ConnectionRate
(/sec)
Number of SYN ACK packets received by the FTP server
per second.
FtpAuthFailed Total number of error replies received by the FTP client.
FTPGets Total number of client Get requests.
FTPPuts Total number of client Put requests.
SuccFTPGets Total number of successful Get requests (data has been
successfully transferred from server to client).
SuccFTPPuts Total number of successful Put requests (data has been
successfully transferred from client to server).
SMTPHelos Total number of HELO packets transmitted.
SMTPHeloRate(/sec) Number of HELO packets transmitted per second.
SMTPHeloLatency
(msec)
Interval between transmitting a HELO packet and
MinSMTPHelo
Latency(msec)
Minimum SMTP Helo Latency in milliseconds.
MaxSMTPHelo
Latency(msec)
Maximum SMTP Helo Latency in milliseconds.
SuccSMTPHelos Total number of successful HELO replies received.
SuccSMTPHelo
Rate(/sec)
Number of successful HELO replies received per second.
SMTPMailFroms Total number of Mail From packets transmitted.
SMTPMailFromRate(
/sec)
Number of Mail From packets transmitted per second.
SMTPMailFrom
Latency(msec)
Interval between transmitting a Mail From packet and
69
MinSMTPMailFromLa
tency(msec)
Minimum SMTP Mail From Latency in milliseconds.
MaxSMTPMailFromL
atency(msec)
Maximum SMTP Mail From Latency in milliseconds.
SuccSMTPMail
Froms
Total number of successful Mail From replies received.
SuccSMTPMailFromR
ate(/sec)
Number of successful Mail From replies received per
second.
SMTPRcptTos Total number of Rcpt To packets transmitted.
SMTPRcptToRate
(/sec)
Number of Rcpt To packets transmitted per second.
SMTPRcptTo
Latency(msec)
Interval between transmitting a Rcpt To packet and
MinSMTPRcptTo
Latency(msec)
Minimum SMTP Rcpt To Latency in milliseconds.
MaxSMTPRcptTo
Latency(msec)
Maximum SMTP Rcpt To Latency in milliseconds.
SuccSMTPRcptTos Total number of successful RcptTo replies received.
SuccSMTPRcptTo
Rate(/sec)
Number of successful Rcpt To replies received per second.
SMTPDatas Total number of Data packets transmitted.
SMTPDataRate(/sec) Number of Data packets transmitted per second.
SMTPDataLatency
(msec)
Interval between transmitting a Data packet and
MinSMTPData
Latency(msec)
Minimum SMTP Data Latency in milliseconds.
MaxSMTPData
Latency(msec)
Maximum SMTP Data Latency in milliseconds.
70
POP3 Performance Counters
SuccSMTPDatas Total number of successful Data replies received.
SuccSMTPDataRate
(/sec)
Number of successful Data replies received per second.
POP3Users Total number of Pop3 User command packets
transmitted.
POP3UserRate(/sec) Number of Pop3 User command packets transmitted per
second.
POP3UserLatency
(msec)
Interval between transmitting a Pop3 User command
packet and receiving a response in milliseconds.
MinPOP3User
Latency(msec)
Minimum POP3UserLatency in milliseconds.
MaxPOP3User
Latency(msec)
Maximum POP3UserLatency in milliseconds.
SuccPOP3Users Total number of successful Pop3 User replies received.
SuccPOP3UserRate
(/sec)
Number of successful Pop3 User replies received per
second.
POP3Passes Total number of Pop3 Pass command packets
transmitted.
POP3PassRate(/sec) Number of Pop3 Pass command packets transmitted per
second.
POP3PassLatency
(msec)
Interval between transmitting a Pop3 Pass packet and
MinPOP3Pass
Latency(msec)
Minimum POP3PassLatency in milliseconds.
MaxPOP3Pass
Latency(msec)
Maximum POP3PassLatency in milliseconds.
SuccPOP3Passes Total number of successful Pop3 Pass replies received.
71
SuccPOP3PassRate
(/sec)
Number of successful Pop3 Pass replies received per
second.
POP3Stats Total number of Pop3 Stat command packets sent.
POP3StatRate(/sec) Number of Pop3 Stat command packets transmitted per
second.
POP3StatLatency
(msec)
Interval between transmitting a Pop3 Stat packet and
MinPOP3Stat
Latency(msec)
Minimum POP3StartLatency in milliseconds.
MaxPOP3Stat
Latency(msec)
Maximum POP3StartLatency in milliseconds.
SuccPOP3Stats Total number of successful Pop3 Stat replies received.
SuccPOP3StatRate
(/sec)
Number of successful Pop3 Stat replies received per
second.
POP3Lists Total number of Pop3 List command packets transmitted.
POP3ListRate(/sec) Number of Pop3 List command packets transmitted per
second.
POP3ListLatency
(msec)
Interval between transmitting a Pop3 List packet and
MinPOP3List
Latency(msec)
Minimum POP3ListLatency in milliseconds.
MaxPOP3List
Latency(msec)
Maximum POP3ListLatency in milliseconds.
SuccPOP3Lists Total number of successful Pop3Lists received.
SuccPOP3ListRate
(/sec)
Number of successful Pop3Lists received per second.
POP3Retrs Total number of Pop3 Retr packets transmitted.
POP3RetrRate(/sec) Number of Pop3 Retr packets transmitted per second.
POP3RetrLatency
(msec)
Interval between transmitting a Pop3 Retr packet and
72
DNS Performance Counters
MinPOP3Retr
Latency(msec)
Minimum POP3RetrLatency in milliseconds.
MaxPOP3Retr
Latency(msec)
Maximum POP3RetrLatency in milliseconds.
SuccPOP3Retrs Total number of successful Pop3Retrs received.
SuccPOP3RetrRate
(/sec)
Number of successful Pop3Retrs received per second.
SuccPrimaryDNS
Request
Total number of Successful DNS requests made to the
Primary DNS server.
SuccSecondaryDNSR
equest
Total number of Successful DNS requests made to the
Secondary DNS server.
SuccDNSData
RequestRate(/sec)
Number of Successful DNS Request packets transmitted
per second.
PrimaryDNSFailure Total number of DNS requests failures received from the
Primary DNS server.
PrimaryDNSRequest Total number of DNS requests made to the Primary DNS
server.
SecondaryDNS
Failure
Total number of DNS requests failures received from the
Secondary DNS server.
SecondaryDNS
Request
Total number of DNS requests made to the Secondary
DNS server.
MinDNSData
Latency
Minimum DNS Data Latency in milliseconds.
MaxDNSData
Latency
Maximum DNS Data Latency in milliseconds.
CurDNSData
Latency
Interval between sending a DNS request packet and
73
Attacks Performance Counters
DNSDataRequest
Rate(/sec)
Number of DNS Request packets transmitted per second.
NoOf
ReTransmission
Total number of DNS Request packets received
NoOfAnswers Total number of Answers to the DNS Request packets.
Attacks Total number of attack packets transmitted (All Attacks)
AttackRate(/sec) Number of attack packets transmitted per second (ARP,
Land, Ping, SYN, and Smurf)
Havoc Flood Number of Havoc packets generated (Stacheldraht only)
Icmp Flood Number of ICMP attack packets generated (TFN, TFN2K,
& Stacheldraht)
Mix Flood Number of Mix packets generated (TFN2K only)
Mstream Flood Number of Mstream packets generated (Stacheldraht
only)
Null Flood Number of Null packets generated (Stacheldraht only)
Smurf Flood Number of Smurf packets generated (TFN, TFN2K, &
Stacheldraht)
Syn Flood Number of SYN packets generated (TFN, TFN2K, &
Stacheldraht)
Targa Flood Number of Targa packets generated (TFN2K only)
Udp Flood Number of UDP packets generated (All DDoS Attacks
only)
74
75
7
Windows Resource Monitoring
The Windows Resources monitor displays statistics about the Windows
resource usage during the load test run. The Windows measurements
correspond to the built-in counters available from the Windows
Performance monitor.
Configuring the Windows Resources Monitor on page 76
Windows Resources Performance Counters on page 79
Chapter 7 Windows Resource Monitoring
76
Configuring the Windows Resources Monitor
Windows NT and Windows 2000 measurements correspond to the built-in
counters available from the Windows Performance monitor. You select the
Windows resources you want to monitor using the Windows Resources
Choose Measurements page.
To monitor Windows resources, we recommend using the SiteScope monitor
engine. In this case, SiteScope must be installed on a server. Although
SiteScope is installed on the Utility Server, we recommend installing it on
the same machine as the Controller, or on a dedicated server in order to
reduce load on the Utility Server.
Note: To monitor a remote Windows machine that does not use Windows
domain security, you must authenticate the Performance Center host
machine on the remote Windows machine. To authenticate the
Performance Center host machine, create an account, or change the
password of the account used to log in to the Performance Center host so
that it matches the password and user name used to log in to the remote
monitored Windows machine. When the remote Windows machine
requests another machines resources, it sends the logged-in user name and
password of the machine requesting the resources.
To monitor a Windows NT or 2000 machine through a firewall, use TCP,
port 139.
To configure the Windows Resources monitor:
Profile.
77
Click Monitors.
2 Under System Resource Graphs, select Windows Resources or Windows
Resources (SiteScope), then click Next. The Choose Server page opens.
3 Type the name or IP address of the server whose resources you want to
monitor. In addition:
For Windows Resources, type the user login name and password on the
Windows machine.
For Windows Resources (SiteScope), type the SiteScope machine name
and port (default: 8888), and specify if you are using a Secure HTTP
connection or a SiteScope user account. If you are using a SiteScope
account, fill in the account name or number, and the user name and
password defined to log in to the SiteScope account.
Click Next.
4 The first time you add measurements using the Windows Resources
(SiteScope) monitor you need to configure the NT remote machine
properties. Enter the remote machines configuration information, as
described in Configuring the NT Remote Machine on page 27.
78
5 The Choose Measurements page opens, displaying the available
6 Select the objects, counters, and instances that you want to monitor. The
instance is relevant only if multiple instances of the highlighted counter are
running.
For a list of available performance counters, see Windows Resources
profile.
79
Windows Resources Performance Counters
The following table describes some of the counters that can be measured:
Object Measurement Description
System % Total Processor
Time
The average percentage of time that all the
processors on the system are busy
executing non-idle threads. On a multi-
processor system, if all processors are
always busy, this is 100%, if all processors
are 50% busy this is 50% and if 1/4th of the
processors are 100% busy this is 25%. It
can be viewed as the fraction of the time
spent doing useful work. Each processor is
assigned an Idle thread in the Idle process
which consumes those unproductive
processor cycles not used by any other
threads.
System File Data
Operations/sec
The rate at which the computer issues read
and write operations to file system devices.
This does not include File Control
Operations.
80
Processor % Processor Time
(Windows 2000)
The percentage of time that the processor is
executing a non-idle thread. This counter
was designed as a primary indicator of
processor activity. It is calculated by
measuring the time that the processor
spends executing the thread of the idle
process in each sample interval, and
subtracting that value from 100%. (Each
processor has an idle thread which
consumes cycles when no other threads are
ready to run). It can be viewed as the
percentage of the sample interval spent
doing useful work. This counter displays
the average percentage of busy time
observed during the sample interval. It is
calculated by monitoring the time the
service was inactive, then subtracting that
value from 100%.
System Processor Queue
Length
The instantaneous length of the processor
queue in units of threads. This counter is
always 0 unless you are also monitoring a
thread counter. All processors use a single
queue in which threads wait for processor
cycles. This length does not include the
threads that are currently executing. A
sustained processor queue length greater
than two generally indicates processor
congestion. This is an instantaneous count,
not an average over the time interval.
Memory Page Faults/sec This is a count of the page faults in the
processor. A page fault occurs when a
process refers to a virtual memory page
that is not in its Working Set in the main
memory. A page fault will not cause the
page to be fetched from disk if that page is
on the standby list (and hence already in
main memory), or if it is in use by another
process with which the page is shared.
81
PhysicalDisk % Disk Time The percentage of elapsed time that the
selected disk drive is busy servicing read or
write requests.
Memory Pool Nonpaged
Bytes
The number of bytes in the nonpaged pool,
a system memory area where space is
acquired by operating system components
as they accomplish their appointed tasks.
Nonpaged pool pages cannot be paged out
to the paging file. They remain in main
memory as long as they are allocated.
Memory Pages/sec The number of pages read from the disk or
written to the disk to resolve memory
references to pages that were not in
memory at the time of the reference. This
is the sum of Pages Input/sec and Pages
Output/sec. This counter includes paging
traffic on behalf of the system cache to
access file data for applications. This value
also includes the pages to/from non-cached
mapped memory files. This is the primary
counter to observe if you are concerned
about excessive memory pressure (that is,
thrashing), and the excessive paging that
may result.
System Total
Interrupts/sec
The rate at which the computer is receiving
and servicing hardware interrupts. The
devices that can generate interrupts are the
system timer, the mouse, data
communication lines, network interface
cards, and other peripheral devices. This
counter provides an indication of how
busy these devices are on a computer-wide
basis. See also Processor:Interrupts/sec.
82
Objects Threads The number of threads in the computer at
the time of data collection. Notice that this
is an instantaneous count, not an average
over the time interval. A thread is the basic
executable entity that can execute
instructions in a processor.
Process Private Bytes The current number of bytes that the
process has allocated that cannot be shared
with other processes.
83
8
SNMP Resource Monitoring
The SNMP Resources monitor is available for monitoring any machine that
runs an SNMP agent, using the Simple Network Management Protocol
(SNMP). SNMP monitoring is platform independent.
Configuring the SNMP Resources Monitor on page 83
Configuring the SiteScope SNMP by MIB Monitor on page 87
Configuring the SNMP Resources Monitor
You select the SNMP resources you want to monitor using the SNMP Choose
Measurements page.
To monitor SNMP resources, we recommend using the SiteScope monitor
engine. In this case, SiteScope must be installed on a server. Although
SiteScope is installed on the Utility Server, you should install it on the same
machine as the Controller, or on a dedicated server in order to reduce load
on the Utility Server.
Note: To monitor SNMP resources through a firewall, use ports 161 or 162.
To configure the SNMP Resources monitor:
Chapter 8 SNMP Resource Monitoring
84
Profile.
Click Monitors.
2 Under System Resource Graphs, select an SNMP monitor:
SNMP (SiteScope) - recommended
SNMP
Click Next. The Choose Server page opens.
monitor. Performance Center connects to default SNMP port 161. To use a
different port number, specify the machine name in the following format:
In addition:
For SNMP (SiteScope), type the SiteScope machine name and port
(default: 8888), and specify if you are using a Secure HTTP connection or
a SiteScope user account. If you are using a SiteScope account, fill in the
account name or number, and the user name and password defined to
log in to the SiteScope account.
For SNMP object nodes with ten or more levels, click Concatenate SNMP
levels to display the sub-levels as a single string, separated by periods (.).
By default, the Choose Measurements page displays all nodes of the
SNMP objects in a tree-level hierarchy.
85
To improve the level of measurement information for the SNMP monitor,
click Show String Item. This lists measurements with string values (in
addition to measurements with numeric values), and displays the string
value as an identifying part of the measurement name.
In the following example, the string value of ProcessName (sched) is
displayed in addition to its instance ID (0):
Click Next.
4 The first time you add measurements using the SNMP (SiteScope) monitor
you need to configure the monitor properties. Enter the monitor properties,
as described in Configuring the SiteScope SNMP by MIB Monitor on
page 87.
86
Note: The SNMP monitor can only monitor up to 25 measurements.
profile.
You can modify the list of resources that you want to monitor at any point
during the load test. Note that a load test does not have to be active in order
for you to monitor the resources on a remote machine.
Note: It may take some time (approximately 2-3 minutes) to save
measurements for the SNMP Resources monitor.
87
Configuring the SiteScope SNMP by MIB Monitor
The first time you add a measurement to the SNMP (SiteScope) monitor, you
need to configure the monitor properties. After you have configured the
monitor properties, you select the measurements to monitor on the
machine.
To configure the SNMP by MIB monitor:
1 After you have selected the SNMP (SiteScope) monitor, and entered the
monitor information, click Next. The Configuring SNMP by MIB Monitor
page opens.
88
2 Type values for the monitor properties:
Server. Type the name of the server you want to monitor.
MIB File. Select the MIB file which contains the objects you are interested
in monitoring. If you select a specific MIB file, then only the objects
described in that MIB file are displayed. If you select All MIBs, then all
objects retrieved from the agent during the MIB traversal will be
displayed.
If no MIB information is available for an object, it is still displayed, but
with no textual name or description. To make this monitor aware of new
or additional MIBs, simply place new MIB files in the
SiteScope/templates.mib directory.
SNMP Version. Select the version of SNMP to use when connecting.
V1/V2 Community. Type the community string (valid only for version 1
or 2 connections).
SNMP V3 Authentication Type. Select the type of authentication to use
for version 3 connections.
SNMP V3 Username. Type the user name for version 3 connections.
SNMP V3 Authentication Password. Type the authentication password to
use for version 3 connections.
SNMP V3 Privacy Password. Type the privacy password if DES privacy
encryption is desired for version 3 connections. Leave blank if you do not
want privacy.
SNMP V3 Context Engine ID. Type a hexadecimal string representing the
Context Engine ID to use for this connection. This is applicable for SNMP
V3 only.
SNMP V3 Context Name. Type the Context Name to use for this
connection. This is applicable for SNMP V3 only.
Timeout. Type the total time, in seconds, that SiteScope should wait for
all SNMP requests (including retries) to complete. The default is 5
seconds.
Retries. Type the number of times each SNMP GET request should be
retried before SiteScope considers the request to have failed. The default
is 1 retry.
89
Port. Type the port to use when requesting data from the SNMP agent.
The default of 161 is the port on which an SNMP agent will typically be
listening.
Update every. Type how frequently the monitor should read the server
statistics. The list to the right of the text box lets you specify time
increments of seconds, minutes, hours, or days. You must specify a time
increment of at least 15 seconds. The default is 30 seconds.
3 Click Next. Add the measurements you want to monitor as described in
step 5 of Configuring the SNMP Resources Monitor on page 83.
90
91
9
SiteScope Resource Monitoring
The SiteScope Resources monitor graph shows the SiteScope resources
measured during the load test.
Note: You can use SiteScope 8.8 or later to monitor multiple instances of the
same process.
The SiteScope monitor can measure server, network, and processor
performance counters. SiteScope can also be used to monitor SAP CCMS
(Computer Center Management System) and SNMP server resources. For
detailed information on all the performance counters that SiteScope can
monitor, see the relevant SiteScope documentation.
Before Setting Up the SiteScope Monitor on page 91
Configuring the SiteScope Monitor on page 92
Before Setting Up the SiteScope Monitor
Before you configure the SiteScope monitor:
Make sure that SiteScope has been installed on a server. Although you
can install SiteScope on the Performance Center host machine, we
recommend that you install it on a dedicated server.
Chapter 9 SiteScope Resource Monitoring
92
On the machine where SiteScope is installed, configure SiteScope to
monitor the required servers. When you assign a name to a monitor,
include the server name in the monitor name. This avoids any confusion
as to which host the monitor belongs. Note that the monitor title cannot
include "\" (for example, CPU Utilization on \\superman).
Note: SiteScopes default sampling rate is 10 minutes, and its minimum
rate is 15 seconds.
Verify that SiteScope is collecting the required data from the servers it is
monitoring. From the SiteScope panel, select the monitor group polling
the server machines, and check that the monitor displays a list of server
measurements in the Status column.
Configuring the SiteScope Monitor
You select the SiteScope resources you want to monitor using the SiteScope
Choose Measurements page.
To configure the SiteScope monitor:
Profile.
Click Monitors.
93
2 Under System Resource Graphs, select SiteScope, and click Next. The
Choose Server page opens.
3 Type the SiteScope Server information as follows:
Server Name. Type the SiteScope server name.
Port. Type the SiteScope port (default: 8888).
Use HTTPS. Specify if you are using a Secure HTTP connection.
Use Account. By default, Performance Center connects to SiteScope using
the default SiteScope Administrator user. If the default SiteScope
Administrator password was changed, or if you want to connect using a
different SiteScope user, select this option. Type the account name, and
the user name and password defined to log in to the relevant SiteScope
account.
Note: The administrator account name is administrator. For details about
how to identify other SiteScope account names, see Identifying the
SiteScope Account Name on page 94.
94
profile.
Note: During configuration of the SiteScope monitor, after selecting the
measurements to monitor and clicking Save, instead of saving the monitor,
Performance Center might go back to the Choose Monitor page. To prevent
this from happening, you need to increase the value of the
AspMaxRequestEntityAllowed property in the IIS metabase. For more
information on troubleshooting this issue, see Cannot Save
Measurements on page 417.
Identifying the SiteScope Account Name
This section describes how to identify the Sitescope account name for a
particular SiteScope user.
To identify a SiteScope account name:
1 Log on to SiteScope as follows: <SiteScope server>:8888/sitescope.
95
2 Click the Preferences button, and then select the Users link.
3 The User Profiles page shows a table that displays the Login URL associated
with each user name. Inside the Login URL you can identify the account
name. In the above example, the account name for the user, TestUSer is
login1.
96
97
10
Server Resource Monitoring
The Server Resources monitor shows the resources of monitors used on
remote Windows and UNIX servers measured during the load test run,
enabling you to determine the impact of Vuser load on the various server
resources.
Setting Up the Server Resources Monitor on page 98
Configuring the Server Resources Monitor on page 98
Server Resources Performance Counters on page 100
Chapter 10 Server Resource Monitoring
98
Setting Up the Server Resources Monitor
To obtain Server Resources data, you need to enable the Server Resources
online monitor (from the Performance Center host) before running the load
test, and indicate which statistics and measurements you want to monitor.
SiteScope must be installed on a server. Although SiteScope is installed on
the Utility Server, we recommend installing it on the same machine as the
Controller, or on a dedicated server in order to reduce load on the Utility
Server.
Configuring the Server Resources Monitor
The Server Resources monitor includes the following monitors:
CPU Monitor. Monitors CPU usage.
Disk Space Monitor. Monitors disk space.
Memory Monitor. Monitors pages per second and percentage of virtual
memory used.
Service Monitor. Verifies that specific processes are listed as running and
checks CPU usage.
To configure the Server Resource monitor:
Profile.
99
Click Monitors.
2 Under System Resource Graphs, select Server Resources (SiteScope) and
click Next. The Choose Server page opens.
3 Under Monitored Machine Information, type the name or IP address of the
server whose resources you want to monitor, and the platform on which the
server runs.
Under SiteScope Server Information, type the SiteScope machine name and
port (default: 8888), and specify if you are using a Secure HTTP connection
or a SiteScope user account. If you are using a SiteScope account, fill in the
account name or number, and the user name and password defined to log in
to the SiteScope account.
Click Next.
4 The first time you add measurements using the Server Resources
(SiteScope) monitor you need to configure the remote machine properties.
If you are monitoring a machine on an NT platform, the Configuring NT
Remote Machine page opens.
If you are monitoring a machine on a UNIX platform, the Configuring
UNIX Remote Machine page opens.
Type the remote machines configuration information, as described in
Configuring the Remote Machine for SiteScope Monitors on page 26.
100
profile.
Server Resources Performance Counters
The following default measurements are available for the Server Resources
machine:
Monitor Measurements Description
CPU Monitor Utilization Measures CPU utilization.
Disk Space Monitor Disk space Measures the percentage of
disk space used.
101
Memory Monitor MB free Measures the amount of disk
space free, in MB.
Pages/sec Measures the number of
virtual memory pages that
are moved between main
memory and disk storage.
Percent used Measures the percentage of
memory and paging file
space used.
Services Monitor Monitors processes locally or
on remote systems. Can be
used to verify that specific
processes are running.
Monitor Measurements Description
102
103
11
UNIX Resource Monitoring
The UNIX Resources monitor shows the UNIX resources measured during
the load test, and enables you to determine the impact of Vuser load on the
various system resources.
The UNIX measurements include those available by the rstatd daemon:
average load, collision rate, context switch rate, CPU utilization, incoming
packets error rate, incoming packets rate, interrupt rate, outgoing packets
error rate, outgoing packets rate, page-in rate, page-out rate, paging rate,
swap-in rate, swap-out rate, system mode CPU utilization, and user mode
CPU utilization.
Configuring the UNIX Resources Monitor on page 106
UNIX Resources Performance Counters on page 108
Chapter 11 UNIX Resource Monitoring
104
To monitor UNIX resources, you must configure the rstatd daemon. The
rstatd daemon might already be configured, because when a machine
receives an rstatd request, the inetd on that machine activates the rstatd
automatically.
To verify whether the rstatd daemon is already configured:
The rup command reports various machine statistics, including rstatd
configuration. Run the following command to view the machine statistics:
You can also use lr_host_monitor and see if it returns any relevant statistics.
If the command returns meaningful statistics, the rstatd daemon is already
configured and activated. If not, or if you receive an error message, the
rstatd daemon is not configured.
To configure the rstatd daemon:
1 Run the command:
2 Go to /etc/inetd.conf and look for the rstatd row (it begins with the word
rstatd). If it is commented out (with a #), remove the comment directive,
and save the file.
3 From the command line, run:
where inet_pid is the pid of the inetd process. This instructs the inetd to
rescan the /etc/inetd.conf file and register all daemons which are
uncommented, including the rstatd daemon.
>rup host
su root
kill -1 inet_pid
105
4 Run rup again.
If the command still does not indicate that the rstatd daemon is configured,
contact your system administrator.
Note: To monitor a UNIX machine through a firewall, you must run a UNIX
utility called rpcinfo and identify the rstatds port number. By running
rpcinfo -p <hostname>, you will receive a list of all RPC servers registered in
the hosts portmapper, along with the port number. This list will not change
until rstatd is stopped and rerun.
Some firewalls allow you to open an RPC program number instead of a port.
In such cases, open program 100001. If you are prompted to include a
version number, specify versions 3 and 4.
106
Configuring the UNIX Resources Monitor
You select the UNIX resources you want to monitor using the UNIX
Resources Choose Measurements page.
To configure the UNIX Resources monitor:
Profile.
Click Monitors.
2 Under System Resource Graphs, select UNIX Resources and click Next. The
monitor.
107
For a list of available performance counters, see UNIX Resources
profile.
108
UNIX Resources Performance Counters
The following default measurements are available for the UNIX machine:
Average load Average number of processes simultaneously in Ready
state during the last minute
Collision rate Collisions per second detected on the Ethernet
Context switches rate Number of switches between processes or threads, per
second
CPU utilization Percent of time that the CPU is utilized
Disk rate Rate of disk transfers
Incoming packets error
rate
Errors per second while receiving Ethernet packets
Incoming packets rate Incoming Ethernet packets per second
Interrupt rate Number of device interrupts per second
Outgoing packets
errors rate
Errors per second while sending Ethernet packets
Outgoing packets rate Outgoing Ethernet packets per second
Page-in rate Number of pages read to physical memory, per second
Page-out rate Number of pages written to pagefiles and removed
from physical memory, per second
Paging rate Number of pages read to physical memory or written
to pagefiles, per second
Swap-in rate Number of processes being swapped
Swap-out rate Number of processes being swapped
System mode CPU
utilization
Percent of time that the CPU is utilized in system
mode
User mode CPU
utilization
Percent of time that the CPU is utilized in user mode
Part V
Network Monitoring
110
111
12
Introducing Network Monitoring
You can use Network monitoring to determine whether your network is
causing a delay in the load test. You can also determine the problematic
network segment.
Network configuration is a primary factor in the performance of
applications. A poorly designed network can slow client activity to
unacceptable levels.
In a true Web or client/server system, there are many network segments. A
single network segment with poor performance can affect the entire system.
The following diagram shows a typical network. To go from the server
machine to the Vuser machine, data must travel over several segments.
To measure network performance, the Network monitor sends packets of
data across the network. When a packet returns, the monitor calculates the
time it takes for the packet to go to the requested node and return. This time
is the delay which appears in the Network Delay Time graph.
Using the online Network Delay Time graph, you can locate the network-
related problem so that it can be fixed.
Chapter 12 Introducing Network Monitoring
112
Note: The delays from the source machine to each of the nodes are
measured concurrently, yet independently. It is therefore possible that the
delay from the source machine to one of the nodes could be greater than the
delay for the complete path between the source and destination machines.
113
13
Network Delay Monitoring
You can use Network monitoring to determine whether your network is
causing a delay in the load test. You can also determine the problematic
network segment.
Network Monitoring from a UNIX Source Machine on page 113
Configuring the Network Delay Time Monitor on page 117
Network Monitoring over a Firewall on page 121
Note: To run the Network monitor, you must have administrator privileges
on the Windows source machine (unless you are using the ICMP protocol).
Network Monitoring from a UNIX Source Machine
You can run the Network monitor on UNIX machines, using UDP or ICMP.
Before running the Network monitor from a UNIX source machine:
Configure the UNIX source machine by assigning root permissions to the
merc_webtrace process (see Configuring the UNIX Source Machine on
page 114).
Make the necessary adjustments to either connect to the source machine
through RSH (see Connecting to the UNIX Source Machine Through
RSH on page 115), or through the agent (see Connecting to the UNIX
Source Machine Through the Agent on page 116).
Chapter 13 Network Delay Monitoring
114
Configuring the UNIX Source Machine
When Performance Center is installed locally, assign root permissions to the
merc_webtrace process by adding an s-bit to merc_webtraces permissions
To configure the source machine, where Performance Center is installed
locally:
1 Log in to the source machine as root.
2 Type: cd <Performance Center installation>/bin to change to the bin directory.
3 Type: chown root merc_webtrace to make the root user the owner of the
merc_webtrace file.
4 Type: chmod +s merc_webtrace to add the s-bit to the file permissions.
5 To verify, type ls -l merc_webtrace. The permissions should look like:
-rwsrwsr-x.
When Performance Center is installed on a network, the merc_webtrace
process is on the network, not on the source machine disk. The procedure
described below copies the merc_webtrace file to the local disk, configures
mdrv.dat to recognize the process, and assigns root permissions to
merc_webtrace.
To configure the source machine, where Performance Center is installed on
the network:
1 Copy merc_webtrace from <Performance Center installation>/bin to
anywhere on the local disk of the source machine. For example, to copy the
file to the /local/<Performance Center> directory, type:
Note: All of the source machines that use the same network installation
must copy merc_webtrace to the identical directory path on their local disk
(for example, /local/<Performance Center>), since all of them use the same
mdrv.dat.
cp /net/tools/<Performance Center installation>/bin/merc_webtrace
/local/<Performance Center>
115
2 Add the following line to the <Performance Center
installation>/dat/mdrv.dat file, in the [monitors_server] section:
3 Log in to the source machine as root.
4 Type: cd <Performance Center installation>/bin to change to the bin directory.
5 Type: chown root merc_webtrace to make the root user the owner of the
merc_webtrace file.
6 Type: chmod +s merc_webtrace to add the s-bit to the file permissions.
7 To verify, type ls -l merc_webtrace. The permissions should look like:
-rwsrwsr-x.
Connecting to the UNIX Source Machine Through RSH
If the Performance Center Host is connected to the source machine through
RSH (default connection mode), then you do not need to activate the agent
daemon. Before running the Network monitor the first time, you type an
encrypted user name and password in the Network monitor configuration
file.
To create an encrypted user name and password:
1 On the Performance Center Host machine, type: cd <Performance Center
installation>/bin to change to the bin directory.
Note: In a network or workstation installation, <Performance Center
installation>/bin is the location on the network in which you installed the
Performance Center setup files.
2 Run CryptonApp.exe.
3 Type your RSH user name and password, separated by a vertical bar symbol.
For example, myname|mypw.
4 Copy the encrypted string to the clipboard (highlight the string and click
CTRL+C).
ExtCmdLine=-merc_webtrace_path /local/xxx
116
5 Add the following line to the <Performance Center
installation>/dat/monitors/ndm.cfg file, in the [hosts] section:
6 Close and open the current load test. Performance Center will read the
updated configuration file and recognize the source machine for
monitoring.
Connecting to the UNIX Source Machine Through the
Agent
If the Performance Center Host is not connected to the source machine
through RSH, then make sure that the agent daemon is active on the source
machine before running the Network monitor.
To activate the agent daemon:
If you are not working in RSH, invoke the agent daemon on the source
machine.
1 Type m_daemon_setup -install from the <Performance Center
installation>/bin directory.
2 Make sure that the agent daemon is running whenever you activate the
Network monitor.
3 To stop the Network Delay Monitor agent daemon, type:
Host = <encrypted string copied from clipboard>
m_daemon_setup - remove
117
Configuring the Network Delay Time Monitor
The Network Delay Time monitor shows the delays for the complete path
between the source and destination machines (for example the database
server and Vuser host). The graph maps the delay as a function of the
elapsed load test time.
You configure the Network monitor from the Monitor tab of the Load Tests
page before you begin running a load test. Using the Network Delay
Monitor page, you select the network path you want to monitor and
configure the network delay monitor settings. You can also configure the
network settings from the Network Delay Monitor Properties page.
Note: To enable network monitoring, you must install the Performance
Center agent on the source machine. You do not have to install the
Performance Center agent on the destination machine.
To configure the Network Delay Time monitor:
Profile.
Click Monitors.
118
2 Under Network Graphs, select Network Delay Time and click Next. The
Network Delay Monitor page opens.
3 In the Source Machine box, type the name of the source machine, from
which you want the network path monitoring to begin.
To run the Network Delay Monitor when there are firewalls between the
Performance Center Host machine and the source machine, in the Source
Machine box, type the server name or IP address of the source machine
according to the format:
where source machine local key is the unique key that you chose when
configuring the Performance Center agent on the source machine.
<MI Listener machine>:<source machine local key>
119
For example,
For more information about working with firewalls in Performance Center,
see the section about firewalls in the HP Performance Center System
Configuration and Installation Guide.
4 Select the platform on which the machine runs.
5 In the Destination Machine box, type the name of the machine at the final
destination of the path you want to monitor.
Note: If the destination machine is localhost, type the local machines name
and not localhost.
6 In the Monitor Settings box, select the protocol and type the port number
being used by the network path. The Network monitor supports three
protocols: TCP, UDP, and ICMP. The default protocol is recommended. In
Windows, the default is TCP, and in UNIX, the default is UDP.
7 Select Enable display of network nodes by DNS names to view the DNS
name of each node along the network path, in addition to its IP address.
Note that selecting this option will decrease the speed of the Network
monitor.
8 In the Monitoring Frequency box, select the number of milliseconds the
monitor should wait between receiving a packet and sending out the next
packet. The default value is 3000 milliseconds. If you have a long, steady
load test, you can increase the interval by several seconds.
12.12.12.3:vds
120
9 In the Monitoring Packet Retries box, select the maximum number of
seconds that the monitor should wait for a packet to return before it retries
to send the packet. The default value is 3 seconds. If your network is very
large and loaded (an internet connection with a low capacity), you should
increase the value by several seconds. If you have a small network (such as a
LAN), you can decrease the value.
In addition, select the number of times the Network monitor should try
resending a packet to a node if the packet is not initially returned. The
default value is 0.
10 Click OK to close the Network Delay Monitor page and save your settings.
The name of the source and destination machines appears in the Choose
Server page.
11 Click Add and repeat this procedure for each path you want to monitor.
To delete a machine, select it and click the Delete button.
To edit the Monitor Settings, click the Edit Profile button. The Network
Delay Monitor Properties page opens. You can modify any of the monitor
settings explained in steps 6 through 9.
12 Click Next, then click Add to add the Network Delay Monitor settings to the
monitor profile.
121
Network Monitoring over a Firewall
If you are monitoring a network in which there are firewalls between the
source and the destination machines, you must configure the firewalls to
allow the network data packets to reach their destinations.
If you are using the TCP protocol, the firewall that protects the destination
machine should not block outgoing ICMP_TIMEEXCEEDED packets
(packets that are sent outside the firewall from the machine). In addition,
the firewall protecting the source machine should allow
ICMP_TIMEEXCEEDED packets to enter, as well as TCP packets to exit.
If you are using the ICMP protocol, the destination machines firewall
should not block incoming ICMP_ECHO_REQUEST packets, or outgoing
ICMP_ECHO_REPLY and ICMP_ECHO_TIMEEXCEEDED packets. In
addition, the firewall protecting the source machine should allow
ICMP_ECHO_REPLY and ICMP_ECHO_TIMEEXCEEDED packets to enter,
and ICMP_ECHO_REQUEST packets to exit.
If you are using the UDP protocol, check that the UDP protocol can access
the destination machine from the source machine. The destination
machines firewall should not block outgoing ICMP_DEST_UNREACHABLE
and ICMP_ECHO_TIMEEXCEEDED packets. In addition, the firewall
protecting the source machine should allow ICMP_DEST_UNREACHABLE
and ICMP_ECHO_TIMEEXCEEDED packets to enter.
122
Part VI
Firewall Server Monitoring
124
125
14
Firewall Server Performance Monitoring
To isolate server performance bottlenecks, you can monitor the firewall
server during a load test run.
About the Firewall Server Monitor on page 125
Configuring the Check Point FireWall-1 Server Monitor on page 125
Check Point FireWall-1 Performance Counters on page 127
About the Firewall Server Monitor
The Firewall server online monitor measures the performance of a Firewall
server during load test execution. To obtain performance data, you must
activate the Firewall server monitor (before executing the load test) and
indicate which statistics and measurements you want to monitor.
Configuring the Check Point FireWall-1 Server Monitor
To monitor the Check Point FireWall-1 server, you must select the counters
you want the Check Point FireWall-1 server monitor to measure.
To configure the Check Point FireWall-1 server monitor:
Chapter 14 Firewall Server Performance Monitoring
126
Profile.
Click Monitors.
2 Under Firewalls, select Check Point FireWall-1 and click Next. The Choose
Server page opens.
monitor. Performance Center connects to port 260, the default port for the
Check Point FireWall-1 SNMP agent. If you do not want to connect to the
default port, specify a machine name and port number using the format:
<machine name>:<port number>
For object nodes with ten or more levels, click Concatenate SNMP levels to
display the sub-levels as a single string, separated by periods (.). By default,
the Choose Measurements page displays all nodes of the SNMP objects in a
tree-level hierarchy.
127
To improve the level of measurement information for the SNMP monitor,
click Show String Item. This enable measurements with string values to be
listed (in addition to measurements with numeric values), and enables the
name modifier (which displays the string value as an identifying part of the
measurement name).
In the following example of a measurement using the name modifier, the
string value of ProcessName (sched) is displayed in addition to its instance
ID (0):
For a list of the available performance counters, see Check Point FireWall-1
Note: The Check Point FireWall-1 monitor can only monitor up to 25
measurements.
profile.
Check Point FireWall-1 Performance Counters
The following default counters can be monitored:
fwRejected The number of rejected packets.
128
fwDropped The number of dropped packets.
fwLogged The number of logged packets.
Part VII
Web Server Monitoring
130
131
15
Introducing Web Server Monitoring
Using Performance Centers Web Server Resource monitors, you can monitor
the Apache, Microsoft IIS, iPlanet (SNMP), and iPlanet/Netscape servers
during a load test run and isolate server performance bottlenecks.
About Web Server Resource Monitors on page 131
Monitoring Using a Proxy Server on page 132
About Web Server Resource Monitors
Web Server Resource monitors provide you with information about the
resource usage of the Apache, Microsoft IIS, iPlanet (SNMP), and
iPlanet/Netscape Web servers during load test execution. To obtain this data,
you need to activate the online monitor for the server and specify which
resources you want to measure before executing the load test.
Chapter 15 Introducing Web Server Monitoring
132
Monitoring Using a Proxy Server
To monitor using the Apache and Netscape monitors, there must be a proxy
server between the Performance Center Host and the monitored server. To
enable this, you must define settings in the relevant configuration files, as
follows:
Before defining settings, you need to determine whether you want
Performance Center to obtain proxy settings from your Internet Explorer
connection configuration, or from the proxy settings in the configuration
file.
To have Performance Center read proxy settings from your Internet
Explorer connection:
1 In the Proxy Settings section of the configuration file, assign useProxy a
value of 1.
2 If the proxy requires a user name, password, or domain, type these
parameters on the lines proxyUsername, proxyPassword, and proxyDomain.
To have Performance Center read proxy settings from the configuration
file:
1 In the Proxy Settings section of the configuration file, type the proxy
information on the httpProxy line. Use the format:
For example:
Monitor Configuration File
Apache <Performance Center root folder>\dat\monitors\
apache.cfg
Netscape <Performance Center root
folder>\dat\monitors\Netscape.cfg
[<protocol>=][<scheme>://]<proxy>[:<port>][[<protocol>=][<scheme>://]
<proxy>[:<port>]]
httpProxy=http=http://my_http_proxy:8080 https=https://my_https_proxy:9000
133
2 If the proxy requires a user name, password, or domain, type these
parameters on the lines: proxyUsername, proxyPassword, and
proxyDomain.
To have Performance Center connect directly to the server (any proxy
settings are ignored):
In the Proxy Settings section of the configuration file, assign useProxy a
value of 0.
134
135
16
Apache Resource Monitoring
The Apache monitor displays statistics about the resource usage on the
Apache server during the load test run.
Configuring the Apache Monitor on page 135
Apache Performance Counters on page 137
Configuring the Apache Monitor
To monitor an Apache server you need to know the server statistics
information URL. A simple way to verify the statistics information URL is to
try to view it through the browser.
The URL should be in the following format:
For example:
To configure the Apache monitor:
http://<server name/IP address>:<port number>/server-status?auto
http://stimpy:80/server-status?auto
Chapter 16 Apache Resource Monitoring
136
Profile.
Click Monitors.
2 Under Web Server Resource Graphs, select Apache and click Next. The
monitor.
In the URL box, type the URL, or accept the default URL:
In the Port box, type the Port number (without the server name), or accept
the default port.
(/server-status?auto)
137
Note: The default port number and URL can vary from one server to
another. Contact your Web server administrator.
To monitor an Apache server through a firewall, use the Web server port (by
default, port 80).
For a list of the available performance counters, see Apache Performance
Counters on page 137.
profile.
Apache Performance Counters
The following table describes the measurements and server properties that
can be monitored:
# Busy Servers The number of servers in the Busy state
# Idle Servers The number of servers in the Idle state
138
Apache CPU Usage The percentage of time the CPU is utilized by the
Apache server
Hits/sec The HTTP request rate
KBytes Sent/sec The rate at which data bytes are sent from the Web
server
139
17
Microsoft IIS Monitoring
The Microsoft IIS monitor displays statistics about the resource usage on the
Microsoft IIS server during the load test run.
Configuring the Microsoft IIS Monitor on page 139
Microsoft IIS Performance Counters on page 142
Configuring the Microsoft IIS Monitor
You select measurements to monitor the Microsoft IIS server from the
Microsoft IIS Choose Measurements page.
To configure the IIS server monitor:
Profile.
Click Monitors.
Chapter 17 Microsoft IIS Monitoring
140
2 Under Web Server Resource Graphs, select MS IIS and click Next. The
monitor.
4 Type the user login name and password.
141
6 Select the objects, counters, and instances that you want to monitor.
For a list of the available performance counters, see Microsoft IIS
profile.
Note: To monitor an IIS server through a firewall, use TCP port 139.
142
Microsoft IIS Performance Counters
Web
Service
Bytes Sent/sec The rate at which the data bytes are sent by the
Web service
Web
Service
Bytes
Received/sec
The rate at which the data bytes are received by
the Web service
Web
Service
Get Requests/sec The rate at which HTTP requests using the GET
method are made. Get requests are generally
used for basic file retrievals or image maps,
though they can be used with forms.
Web
Service
Post Requests/sec The rate at which HTTP requests using the POST
method are made. Post requests are generally
used for forms or gateway requests.
Web
Service
Maximum
Connections
The maximum number of simultaneous
connections established with the Web service
Web
Service
Current
Connections
The current number of connections established
with the Web service
Web
Service
Current
NonAnonymous
Users
The number of users that currently have a non-
anonymous connection using the Web service
Web
Service
Not Found
Errors/sec
The rate of errors due to requests that could not
be satisfied by the server because the requested
document could not be found. These are
generally reported to the client as an HTTP 404
error code.
Process Private Bytes The current number of bytes that the process
has allocated that cannot be shared with other
processes.
143
18
iPlanet/Netscape Monitoring
The iPlanet/Netscape monitor displays statistics about the resource usage on
the iPlanet/Netscape server during the load test run.
Configuring the iPlanet/Netscape Monitor on page 144
iPlanet/Netscape Performance Counters on page 147
Chapter 18 iPlanet/Netscape Monitoring
144
Configuring the iPlanet/Netscape Monitor
To monitor an iPlanet/Netscape server, you need to know the
administration server URL. A simple way to verify the administration server
URL, is to try to view it through the browser.
The URL should be in the following format:
for example:
Notes:
In some server configurations, the URL must contain the administration
server name and not the IP address.
In addition, the administration server name may differ from the
iPlanet/Netscape server name.
To monitor an iPlanet/Netscape server through a firewall, use the
iPlanet/Netscape Administration server port. Configure this port during
the server installation process.
To activate the iPlanet/Netscape monitor from the Performance Center
Host:
Profile.
http://<admin_srv_name/IP address>:<port number>/https-<admin_srv_name/
IP address>/bin/sitemon?doit
http://lazarus:12000/https-lazarus.mercury.co.il/bin/sitemon?doit
145
Click Monitors.
2 Under Web Server Resource Graphs, select iPlanet/Netscape and click Next.
The Choose Server page opens.
monitor.
another. Please consult the Web server administrator. In some server
configurations, the URL must contain the administration server name and
not the IP address.
In the URL box, type the URL or accept the default URL,
(/https-<admin_server>/bin/sitemon?doit)
146
In the Port box, type the Port number (without the server name), or
accept the default port.
Type the user login name and password. The user must have
administrator permissions on the server.
5 Select the required measurements.
For a list of the available performance counters, see iPlanet/Netscape
profile.
147
iPlanet/Netscape Performance Counters
can be monitored:
200/sec The rate of successful transactions being processed by the
server
2xx/sec The rate at which the server handles status codes in the 200
to 299 range
302/sec The rate of relocated URLs being processed by the server
304/sec The rate of requests for which the server tells the user to use
a local copy of a URL instead of retrieving a newer version
from the server
to 399 range
401/sec The rate of unauthorized requests handled by the server
403/sec The rate of forbidden URL status codes handled by the server
to 499 range
5xx/sec The rate at which the server handles status codes 500 and
higher
Bad requests/sec The rate at which the server handles bad requests
Bytes sent/sec The rate at which bytes of data are sent from the Web server
Hits/sec The HTTP request rate
xxx/sec The rate of all status codes (2xx-5xx) handled by the server,
excluding timeouts and other errors that did return an HTTP
status code
148
149
19
iPlanet (SNMP) Monitoring
The iPlanet (SNMP) monitor uses the Simple Network Management Protocol
(SNMP) to retrieve iPlanet (SNMP) server statistics.
Configuring the iPlanet (SNMP) Monitor on page 149
iPlanet (SNMP) Performance Counters on page 151
Configuring the iPlanet (SNMP) Monitor
You select measurements to monitor the iPlanet (SNMP) server using the
iPlanet (SNMP) Choose Measurements page.
Note: To monitor a iPlanet (SNMP) server, use port 161 or 162, depending
on the configuration of the agent.
To configure the iPlanet (SNMP) Resources monitor:
Profile.
Chapter 19 iPlanet (SNMP) Monitoring
150
Click Monitors.
2 Under Web Server Resource Graphs, select iPlanet (SNMP) and click Next.
monitor. You need to define the port number if the iPlanet SNMP agent is
running on a different port than the default SNMP port. Specify the
machine name and port number in the format:
display the sub-levels as a single string, separated by periods (.). By default,
the Choose Measurements page displays all nodes of the SNMP objects in a
tree-level hierarchy.
To improve the level of measurement information for the iPlanet SNMP
monitor, click Show String Item. This lists measurements with string values
(in addition to measurements with numeric values), and displays the string
151
For a list of the available performance counters, see iPlanet (SNMP)
Note: The iPlanet (SNMP) monitor can monitor up to 25 measurements at a
time.
profile.
iPlanet (SNMP) Performance Counters
can be monitored:
iwsInstanceTable iPlanet Web Server instances
iwsInstanceEntry iPlanet Web Server instances
iwsInstanceIndex Server instance index
iwsInstanceId Server instance identifier
iwsInstanceVersion Server instance software version
152
iwsInstanceDescription Description of server instance
iwsInstanceOrganization Organization responsible for server instance
iwsInstanceContact Contact information for persons responsible for
server instance
iwsInstanceLocation Location of server instance
iwsInstanceStatus Server instance status
iwsInstanceUptime Server instance uptime
iwsInstanceDeathCount Number of times server instance processes have
died
iwsInstanceRequests Number of requests processed
iwsInstanceInOctets Number of octets received
iwsInstanceOutOctets Number of octets transmitted
iwsInstanceCount2xx Number of 200-level (Successful) responses
issued
iwsInstanceCount3xx Number of 300-level (Redirection) responses
issued
iwsInstanceCount4xx Number of 400-level (Client Error) responses
issued
iwsInstanceCount5xx Number of 500-level (Server Error) responses
issued
iwsInstanceCountOther Number of other (neither 2xx, 3xx, 4xx, nor
5xx) responses issued
iwsInstanceCount200 Number of 200 (OK) responses issued
iwsInstanceCount302 Number of 302 (Moved Temporarily) responses
issued
iwsInstanceCount304 Number of 304 (Not Modified) responses issued
iwsInstanceCount400 Number of 400 (Bad Request) responses issued
iwsInstanceCount401 Number of 401 (Unauthorized) responses issued
153
iwsInstanceCount403 Number of 403 (Forbidden) responses issued
iwsInstanceCount404 Number of 404 (Not Found) responses issued
iwsInstanceCount503 Number of 503 (Unavailable) responses issued
iwsInstanceLoad
1MinuteAverage
System load average for 1 minute
iwsInstanceLoad
5MinuteAverage
System load average for 5 minutes
iwsInstanceLoad
15MinuteAverage
System load average for 15 minutes
iwsInstanceNetwork
InOctets
Number of octets transmitted on the network
per second
iwsInstanceNetwork
OutOctets
Number of octets received on the network per
second
iwsVsTable iPlanet Web Server virtual servers
iwsVsEntry iPlanet Web Server virtual server
iwsVsIndex Virtual server index
iwsVsId Virtual server identifier
iwsVsRequests Number of requests processed
iwsVsInOctets Number of octets received
iwsVsOutOctets Number of octets transmitted
iwsVsCount2xx Number of 200-level (Successful) responses
issued
iwsVsCount3xx Number of 300-level (Redirection) responses
issued
iwsVsCount4xx Number of 400-level (Client Error) responses
issued
iwsVsCount5xx Number of 500-level (Server Error) responses
issued
154
iwsVsCountOther Number of other (neither 2xx, 3xx, 4xx, nor
5xx) responses issued
iwsVsCount200 Number of 200 (OK) responses issued
iwsVsCount302 Number of 302 (Moved Temporarily) responses
issued
iwsVsCount304 Number of 304 (Not Modified) responses issued
iwsVsCount400 Number of 400 (Bad Request) responses issued
iwsVsCount401 Number of 401 (Unauthorized) responses issued
iwsVsCount403 Number of 403 (Forbidden) responses issued
iwsVsCount404 Number of 404 (Not Found) responses issued
iwsVsCount503 Number of 503 (Unavailable) responses issued
iwsProcessTable iPlanet Web Server processes
iwsProcessEntry iPlanet Web Server process
iwsProcessIndex Process index
iwsProcessId Operating system process identifier
iwsProcessThreadCount Number of request processing threads
iwsProcessThreadIdle Number of request processing threads currently
idle
iwsProcessConnection
QueueCount
Number of connections currently in connection
queue
QueuePeak
Largest number of connections that have been
queued simultaneously
QueueMax
Maximum number of connections allowed in
connection queue
QueueTotal
Number of connections that have been accepted
QueueOverflows
Number of connections rejected due to
connection queue overflow
155
iwsProcessKeepalive
Count
Number of connections currently in keepalive
queue
iwsProcessKeepaliveMax Maximum number of connections allowed in
keepalive queue
iwsProcessSizeVirtual Process size in kbytes
iwsProcessSizeResident Process resident size in kbytes
iwsProcessFraction
SystemMemoryUsage
Fraction of process memory in system memory
iwsListenTable iPlanet Web Server listen sockets
iwsListenEntry iPlanet Web Server listen socket
iwsListenIndex Listen socket index
iwsListenId Listen socket identifier
iwsListenAddress Address socket is listening on
iwsListenPort Port socket is listening on
iwsListenSecurity Encryption support
iwsThreadPoolTable iPlanet Web Server thread pools
iwsThreadPoolEntry iPlanet Web Server thread pool
iwsThreadPoolIndex Thread pool index
iwsThreadPoolId Thread pool identifier
iwsThreadPoolCount Number of requests queued
iwsThreadPoolPeak Largest number of requests that have been
queued simultaneously
iwsThreadPoolMax Maximum number of requests allowed in queue
iwsCpuTable iPlanet Web Server CPUs
iwsCpuEntry iPlanet Web Server CPU
iwsCpuIndex CPU index
iwsCpuId CPU identifier
156
iwsCpuIdleTime CPU Idle Time
iwsCpuUserTime CPU User Time
iwsCpuKernelTime CPU Kernel Time
Part VIII
Web Application Server Resource
Monitoring
158
159
20
Introducing Web Application Server
Resource Monitoring
You can monitor a Web application server during a load test run and isolate
application server performance bottlenecks using Performance Centers Web
Application Server Resource monitors.
Web Application Server Resource monitors provide you with information
about the resource usage of the following servers during a load test run:
Ariba
ATG Dynamo
BroadVision
ColdFusion
iPlanet (NAS)
Microsoft ASP
Oracle9iAS HTTP
SilverStream
WebLogic (JMX)
WebSphere
To obtain performance data, you need to activate the online monitor for the
server and specify which resources you want to measure before executing
the load test.
Chapter 20 Introducing Web Application Server Resource Monitoring
160
161
21
Ariba Monitoring
The Ariba monitor displays statistics about the resource usage on the Ariba
server during the load test run.
Configuring the Ariba Monitor on page 161
Ariba Performance Counters on page 164
Configuring the Ariba Monitor
You select the measurements you want to monitor on the Ariba server using
the Ariba Choose Measurements page.
Note: The port you use to monitor an Ariba server through a firewall
depends on the configuration of your server.
To configure the Ariba monitor:
Profile.
Chapter 21 Ariba Monitoring
162
Click Monitors.
2 Under Web Application Server Resource Graphs, select Ariba and click Next.
163
3 Type the server name or IP address in the format:
<server name>:<port number>. For example, merc1:12130.
For a list of the available performance counters, see Ariba Performance
profile.
XML Accessibility Verification
Only browsers that are XML-compatible will allow you to view the
performance XML file.
164
To verify whether the XML file is accessible:
Display the XML file through the browser. The URL should be in the format:
For example,
Note: In some cases, although the browser is XML-compatible, it may still
return the error: The XML page cannot be displayed. In these cases, the XML
file can be accessed by the Ariba performance monitor, although it cannot
be viewed by the browser.
Ariba Performance Counters
Core Server Performance Counters
http://<server name:port number>/metrics?query=getStats
http://merc1:12130/metrics?query=getStats
Requisitions Finished The instantaneous reading of the length of the worker
queue at the moment this metric is obtained. The longer
the worker queue, the more user requests are delayed for
processing
Worker Queue
Length
The instantaneous reading of the length of the worker
queue at the moment this metric is obtained. The longer
the worker queue, the more user requests are delayed for
processing
Concurrent
Connections
The instantaneous reading of the number of concurrent
user connections at the moment this metric is obtained
165
Total Connections The cumulative number of concurrent user connections
since Ariba Buyer was started
Total Memory The instantaneous reading of the memory (in KB) being
used by Ariba Buyer at the moment this metric is
obtained
Free Memory The instantaneous reading of the reserved memory (in
bytes) that is not currently in use at the moment this
metric is obtained
Up Time The amount of time (in hours and minutes) that Ariba
Buyer has been running since the previous time it was
started
Number of Threads The instantaneous reading of the number of server
threads in existence at the moment this metric is
obtained
Number of Cached
Objects
The instantaneous reading of the number of Ariba Buyer
objects being held in memory at the moment this metric
is obtained
Average Session
Length
The average length of the user sessions (in seconds) of all
users who logged out since previous sampling time. This
value indicates on average how long a user stays
connected to server
Average Idle Time The average idle time (in seconds) for all the users who
are active since previous sampling time. The idle time is
the period of time between two consecutive user requests
from the same user
Approves The cumulative count of the number of approves that
happened during the sampling period. An Approve
consists of a user approving one Approvable
Submits The cumulative count of the number of Approvals
submitted since previous sampling time
Denies The cumulative count of the number of submitted
Approvals denied since previous sampling time
166
System Related Performance Counters
Object Cache
Accesses
The cumulative count of accesses (both reads and writes)
to the object cache since previous sampling time
Object Cache Hits The cumulative count of accesses to the object cache that
are successful (cache hits) since previous sampling time
Database Response
Time
The average response time (in seconds) to the database
requests since the previous sampling time
Buyer to DB server
Traffic
The cumulative number of bytes that Ariba Buyer sent to
DB server since the previous sampling time
DB to Buyer server
Traffic
The cumulative number of bytes that DB server sent to
Ariba Buyer since the previous sampling time
Database Query
Packets
The average number of packets that Ariba Buyer sent to
DB server since the previous sampling time
Database Response
Packets
The average number of packets that DB server sent to
Ariba Buyer since the previous sampling time
167
22
ATG Dynamo Monitoring
The ATG Dynamo monitor uses SNMP to retrieve ATG Dynamo server
statistics.
Configuring the ATG Dynamo Monitor on page 167
ATG Dynamo Performance Counters on page 170
Configuring the ATG Dynamo Monitor
You select the measurements you want to monitor on the ATG Dynamo
server using the ATG Dynamo Choose Measurements page.
To configure the ATG Dynamo server monitor:
Profile.
Click Monitors.
Chapter 22 ATG Dynamo Monitoring
168
2 Under Web Application Server Resource Graphs, select ATG Dynamo and
3 Type the server name or IP address of the machine you want to monitor.
Note that you need to define the port number if the ATG SNMP agent is
running on a different port than the default ATG SNMP port 8870. Specify a
machine name and port number using the format:
For example,
display the sub-levels as a single string, separated by periods (.). In the
default view, the Choose Measurements page displays all nodes of the SNMP
objects in a tree-level hierarchy.
To improve the level of measurement information for the ATG Dynamo
<server name:port number>
merc1:12130
169
For a list of the available performance counters, see ATG Dynamo
Note: The ATG Dynamo monitor can only monitor up to 25 measurements.
170
profile.
ATG Dynamo Performance Counters
The following tables describe the measurements that can be monitored:
d3System
d3LoadManagement
sysTotalMem The total amount of memory currently available for
allocating objects, measured in bytes
sysFreeMem An approximation of the total amount of memory
currently available for future allocated objects,
measured in bytes
sysNumInfoMsgs The number of system global info messages written
sysNumWarningMsgs The number of system global warning messages written
sysNumErrorMsgs The number of system global error messages written
lmIsManager True if the Dynamo is running a load manager
lmManagerIndex Returns the Dynamo's offset into the list of load
managing entities
lmIsPrimaryManager True if the load manager is an acting primary manager
lmServicingCMs True if the load manager has serviced any connection
module requests in the amount of time set as the
connection module polling interval
lmCMLDRPPort The port of the connection module agent
lmIndex A unique value for each managed entity
lmSNMPPort The port for the entry's SNMP agent
171
d3SessionTracking
d3DRPServer
lmProbability The probability that the entry will be given a new
session
lmNewSessions Indicates whether or not the entry is accepting new
sessions, or if the load manager is allowing new
sessions to be sent to the entry. This value is inclusive
of any override indicated by lmNewSessionOverride
lmNewSessionOverride The override set for whether or not a server is
accepting new sessions
stCreatedSessionCnt The number of created sessions
stValidSessionCnt The number of valid sessions
stRestoredSessionCnt The number of sessions migrated to the server
stDictionaryServerStatus d3Session Tracking
drpPort The port of the DRP server
drpTotalReqsServed Total number of DRP requests serviced
drpTotalReqTime Total service time in msecs for all DRP requests
drpAvgReqTime Average service time in msecs for each DRP request
drpNewessions True if the Dynamo is accepting new sessions
172
d3DBConnPooling
dbPoolsEntry A pooling service entry containing information
about the pool configuration and current status
dbIndex A unique value for each pooling service
dbPoolID The name of the DB connection pool service
dbMinConn The minimum number of connections pooled
dbMaxConn The maximum number of connections pooled
dbMaxFreeConn The maximum number of free pooled connections
at a time
dbBlocking Indicates whether or not the pool is to block out
check outs
dbConnOut Returns the number of connections checked out
dbFreeResources Returns the number of free connections in the pool.
This number refers to connections actually created
that are not currently checked out. It does not
include how many more connections are allowed to
be created as set by the maximum number of
connections allowed in the pool
dbTotalResources Returns the number of total connections in the
pool. This number refers to connections actually
created and is not an indication of how many more
connections may be created and used in the pool
173
23
BroadVision Monitoring
The BroadVision monitor displays performance statistics for all the
servers/services available on the BroadVision application during a load test
run.
Configuring the BroadVision Monitor on page 174
BroadVision Servers and Services on page 177
BroadVision Performance Counters on page 179
Chapter 23 BroadVision Monitoring
174
Configuring the BroadVision Monitor
To monitor a BroadVision server, you must grant the client permission to
invoke or launch services on the server. You select the measurements you
want to monitor on the BroadVision server using the BroadVision Choose
Measurements page.
Note: The port you use to monitor a BroadVision server through a firewall
To grant permission for a BroadVision server:
Use the Iona Technologies (Orbix) command for setting user and access
permission on a load generator machine:
If you experience problems connecting to the BroadVision monitor, you
may need to redefine the permissions to "all."
To invoke permission for all, type the following command at the
BroadVision server command prompt:
To launch permission for all, type the following command at the
BroadVision server command prompt:
chmodit [-h <host>] [-v] { <server> | -a <dir> }
{i{+,-}{user,group} | l{+,-}{user,group} }
# chmodit <server> i+all
# chmodit <server> l+all
175
Alternatively, set ORBIX_ACL. Setting ORBIX_ACL=i+all l+all in the
BroadVision/Orbix configuration file gives permission to all.
In addition, to monitor a BroadVision server, you need to have JDK 1.2 or
higher installed on the Performance Center Host machine.
You can install JDK 1.2 by following the download and installation
instructions at the following Web site: http://java.sun.com/products/jdk/1.2/
Before activating the monitor, make sure that your Java environment is
configured properly.
To configure your Java environment:
1 Open the Windows Registry.
2 The registry should contain the correct path to the Java executable
(java.exe) under the JDK 1.2 installation directory. Verify the following
registry key:
3 The registry should contain the correct path to the Java run-time
environment (JRE) under the JRE 1.2 installation directory. Verify the
following registry key:
To configure the BroadVision online monitor:
Profile.
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\App
Paths\java.exe
HKEY_LOCAL_MACHINE\SOFTWARE\JavaSoft\Java Runtime
Environment\1.2\JavaHome
176
Click Monitors.
2 Under Web Application Server Resource Graphs, select either
BroadVision(4.5-5.4) or BroadVision (5.5-6.X) and click Next. The relevant
3 Type the BroadVision server name or IP address with the port number
<server name>:<port number>. For example, dnsqa:1221.
177
For a list of the available performance counters, see BroadVision Servers
and Services on page 177.
profile.
BroadVision Servers and Services
The following table describes the servers and services that can be monitored:
Server
Multiple
Instances
Description
adm_srv No One-To-One user administration server. There
must be one.
alert_srv No Alert server handles direct IDL function calls to
the Alert system.
178
bvconf_srv No One-To-One configuration management server.
There must be one.
cmsdb Yes Visitor management database server.
cntdb Yes Content database server.
deliv_smtp_d Yes Notification delivery server for email type
messages. Each instance of this server must have
its own ID, numbered sequentially starting
with 1.
deliv_comp_d No Notification delivery completion processor.
extdbacc Yes External database accessor. You need at least one
for each external data source.
genericdb No Generic database accessor handles content query
requests from applications, when specifically
called from the application. This is also used by
the One-To-One Command Center.
hostmgr Yes Defines a host manager process for each machine
that participates in One-To-One, but doesn't run
any One-To-One servers. For example, you need a
hostmgr on a machine that runs only servers. You
don't need a separate hostmgr on a machine that
already has one of the servers in this list.
g1_ofbe_srv No Order fulfillment back-end server.
g1_ofdb Yes Order fulfillment database server.
g1_om_srv No Order management server.
pmtassign_d No The payment archiving daemon routes payment
records to the archives by periodically checking
the invoices table, looking for records with
completed payment transactions, then moving
those records into an archive table.
Server
Multiple
Instances
Description
179
BroadVision Performance Counters
Performance counters for each server or service are divided into logical
groups according to the service type.
The following section describes all the available counters under each group.
Note that the same group can have a different number of counters,
depending on the service.
Counter groups:
BV_DB_STAT
BV_SRV_CTRL
BV_SRV_STAT
NS_STAT
BV_CACHE_STAT
JS_SCRIPT_CTRL
JS_SCRIPT_STAT
pmthdlr_d Yes For each payment processing method, you need
one or more authorization daemons to
periodically acquire the authorization when a
request is made.
pmtsettle_d Yes Payment settlement daemon periodically checks
the database for orders of the associated payment
processing method that need to be settled, then
authorizes the transactions.
sched_poll_d No Notification schedule poller scans the database
tables to determine when a notification must be
run.
sched_srv Yes Notification schedule server runs the scripts that
generate the visitor notification messages.
Server
Multiple
Instances
Description
180
BV_DB_STAT
The database accessor processes have additional statistics available from the
BV_DB_STAT memory block. These statistics provide information about
database accesses, including the count of selects, updates, inserts, deletes,
and stored procedure executions.
BV_SRV_CTRL
BV_SRV_STAT
The display for Interaction Manager processes includes information about
the current count of sessions, connections, idle sessions, threads in use, and
count of CGI requests processed.

DELETE Count of deletes executions
INSERT Count of inserts executions
SELECT Count of selects executions
SPROC Count of stored procedure executions
UPDATE Count of update executions
SHUTDOWN Shut-down trigger. Setting this to non-zero value
causes the server to shut down.
HOST Host machine running the process
ID Instance of the process (of which multiple can be
configured in the bv1to1.conf file), or engine ID of
the Interaction Manager
CGI Current count of CGI requests processed
CONN Current count of connections
181
CPU CPU percentage consumed by this process. If a
process is using most of the CPU time, consider
moving it to another host, or creating an additional
process, possibly running on another machine.
Both of these specifications are done in the
bv1to1.conf file. The CPU % reported is against a
single processor. If a server is taking up a whole CPU
on a 4 processor machine, this statistic will report
100%, while the Windows Task Manager will report
25%. The value reported by this statistic is
consistent with "% Processor Time" on the Windows
Performance Monitor
GROUP Process group (which is defined in the bv1to1.conf
file), or Interaction Manager application name.
STIME Start time of server. The start times should be
relatively close. Later times might be an indication
that a server crashed and was automatically
restarted.
IDL Total count of IDL requests received, not including
those to the monitor.
IdlQ
JOB
LWP Number of light-weight processes (threads).
RSS Resident memory size of server process (in
kilobytes).
STIME System start time.
SESS Current count of sessions.
SYS Accumulated system mode CPU time (seconds).
THR Current count of threads.
182
NS_STAT
The NS process displays the namespace for the current One-To-One
environment, and optionally can update objects in a name space.

BV_CACHE_STAT
Monitors the request cache status.
The available counters for each request are:

USR Accumulated user mode CPU time (seconds).
VSZ Virtual memory size of server process (in kilobytes).
If a process is growing in size, it probably has a
memory leak. If it is an Interaction Manager
process, the culprit is most likely a component or
dynamic object (though Interaction Manager
servers do grow and shrink from garbage collection
during normal use).
Bind Count of bind() CORBA requests served
List Count of list() CORBA requests served
New Count of new NamingContext objects created
Rebnd Count of rebind() CORBA requests served
Rsolv Count of resolve() CORBA requests served
Unbnd Count of unbind() CORBA requests served
CNT- Request_Name-HIT Count of requests found in the cache.
CNT- Request_Name-MAX Maximum size of the cache in bytes.
CNT- Request_Name-SWAP Count of items that got swapped out of the
cache.
CNT- Request_Name-MISS Count of requests that were not in the cache.
CNT- Request_Name-SIZE Count of items currently in the cache.
183
Cache Metrics
Cache metrics are available for the following items:
AD Contains information about advertisements
(current cache size, configured cache size, count of
hits, count of misses, count of replacements).
ALERTSCHED Notification schedules are defined in the
BV_ALERTSCHED and BV_MSGSCHED tables. They
are defined by the One-To-One Command Center
user or by an application.
CATEGORY_CONTENT Contains information about the category (current
cache size, configured cache size, count of hits,
count of misses, count of replacements).
DISCUSSION The One-To-One discussion groups provide
moderated system of messages and threads of
messages aligned to a particular topic. Use the
Discussion group interfaces for creating, retrieving
and deleting individual messages in a discussion
group. To create, delete, or retrieve discussion
groups, use the generic content management API.
The BV_DiscussionDB object provides access to the
threads and messages in the discussion group
database.
EXT_FIN_PRODUCT Contains information about the external product
(current cache size, configured cache size, count of
hits, count of misses, count of replacements).
EDITORIAL Using the Editorials content module, you can point
cast and community cast personalized editorial
content, and sell published text on your One-To-
One site. You can solicit editorial content, such as
investment reports and weekly columns, from
outside authors and publishers, and create your
own articles, reviews, reports, and other informative
media. In addition to text, you can use images,
sounds, music, and video presentations as editorial
content.
INCENTIVE Contains sales incentives.
184
MSGSCHED Contains the specifications of visitor-message jobs.
Notification schedules are defined in the
BV_ALERTSCHED and BV_MSGSCHED tables. They
are defined by the One-To-One Command Center
user or by an application.
MSGSCRIPT Contains the descriptions of the JavaScripts that
generate visitor messages and alert messages.
Contains the descriptions of the JavaScripts that
generate targeted messages and alert messages. Use
the Command Center to add message script
information to this table by selecting the Visitor
Messages module in the Notifications group. For
more information, see the Command Center User's
Guide.
PRODUCT BV_PRODUCT contains information about the
products that a visitor can purchase.
QUERY BV_QUERY contains queries.
SCRIPT BV_SCRIPT contains page scripts.
SECURITIES Contains information about securities (current
cache size, configured cache size, count of hits,
count of misses, count of replacements)
TEMPLATE The Templates content module enables you to store
in the content database any BroadVision page
templates used on your One-To-One site.
Combining BroadVision page templates with
BroadVision dynamic objects in the One-To-One
Design Center application is one way for site
developers to create One-To-One Web sites. If your
developers use these page templates, you can use
the Command Center to enter and manage them in
your content database. If your site doesn't use
BroadVision page template, you will not use this
content module.
185
JS_SCRIPT_CTRL
JS_SCRIPT_STAT
CACHE Script cache enabled. Initially 1 (on).
DUMP Reference dump sequence. Increment this to cause
all contexts to dump their references when they
next run their next script. The dump information is
appended to the jsctxdump.<pid> file in the
BVLOG directory.
FLUSH Cache flushing sequence. Increment this to cause
all contexts to flush their script cache when they
next run their next script.
METER Metering enabled. Initially 0 (off).
TRACE Tracing level. Range is BVLOG_LEVEL_MIN to
BVLOG_LEVEL_MAX. Lower level means more
tracing. Initial value is BVLOG_DEBUG.
ALLOC Count of reference allocations
ERROR Count of recoverable errorshow many times the
Error object was set
FAIL Count of failed script runs
JSPPERR Count of preprocessor failuresalso counted in
FAIL
RELEASE Count of reference releases
STOP Count of stopped script runs
SUCC Count of successful script runs
SYNTAX Count of JavaScript compilation failuresalso
counted in FAIL
186
187
24
ColdFusion Monitoring
The ColdFusion monitor displays statistics about the resource usage on the
ColdFusion server during the load test run.
Configuring the ColdFusion Monitor on page 188
ColdFusion Performance Counters on page 190
Chapter 24 ColdFusion Monitoring
188
Configuring the ColdFusion Monitor
You select the measurements you want to monitor on the ColdFusion server
using the ColdFusion Choose Measurements page.
Note: The ColdFusion monitor works through HTTP and supports UNIX
platforms. To monitor the ColdFusion server on Windows platforms, you
can also use the Windows Resource monitor.
To set up the ColdFusion monitor environment:
Copy the <Performance Center installation>\dat\monitors\perfmon.cfm
file into the <ColdFusion Home>\cfide\administrator directory. By default,
the ColdFusion monitor checks for the
<ColdFusion Home>\cfide\administrator\perfmon.cfm file.
Note: The port you use to monitor a ColdFusion server through a firewall
To configure the ColdFusion monitor:
Profile.
189
Click Monitors.
2 Under Web Application Server Resource Graphs, select ColdFusion and click
monitor.
Type the user login name and password.
For a list of the available performance counters, see ColdFusion
profile.
190
ColdFusion Performance Counters
The following table describes the default counters that can be measured:
Avg. Database Time
(msec)
The running average of the amount of time, in
milliseconds, that it takes ColdFusion to process
database requests.
Avg. Queue Time (msec) The running average of the amount of time, in
milliseconds, that requests spent waiting in the
ColdFusion input queue before ColdFusion began
to process the request.
Avg Req Time (msec) The running average of the total amount of time, in
milliseconds, that it takes ColdFusion to process a
request. In addition to general page processing time,
this value includes both queue time and database
processing time.
Bytes In/sec The number of bytes per second sent to the
ColdFusion server.
Bytes Out/sec The number of bytes per second returned by the
ColdFusion server.
Cache Pops Cache pops.
Database Hits/sec This is the number of database hits generated per
second by the ColdFusion server.
Page Hits/sec This is the number of Web pages processed per
second by the ColdFusion server.
Queued Requests The number of requests currently waiting to be
processed by the ColdFusion server.
Running Requests The number of requests currently being actively
processed by the ColdFusion server.
Timed Out Requests The number of requests that timed out due to
inactivity timeouts.
191
25
iPlanet (NAS) Monitoring
The iPlanet (NAS) monitor uses the Simple Network Management Protocol
(SNMP) to retrieve iPlanet (NAS) server statistics.
Configuring the iPlanet (NAS) Monitor on page 196
iPlanet (NAS) Performance Counters on page 198
Chapter 25 iPlanet (NAS) Monitoring
192
Before configuring the iPlanet (NAS) monitor, you need to set up the
application server for SNMP monitoring.
SNMP Summary on page 192
Setting up SNMP Monitoring of the iPlanet Application Server on
page 193
Note:
This section offers a short explanation about setting up SNMP
monitoring of the iPlanet Application Server. It is intended to
supplement the iPlanet documentation. For more information about
SNMP reporting architecture and theory, see the iPlanet documentation.
The instructions below assume that SNMP statistics will be collected on
the standard SNMP port 161.\
SNMP Summary
Solaris has a native SNMP agent, snmpdx, that is started automatically at
boot time by the script /etc/rc3.d/S76snmpdx. This daemon communicates
on the standard SNMP port 161. The port number can be changed with the
-p <port> option.
Planet products are shipped with their own SNMP agents. The architecture is
such that there is one master agent per host, which a network management
station communicates with, and one or more subagents that collect data
from various iPlanet products and forward statistics to the master agent. The
master agent also defaults to communicating on port 161.
To run both the Solaris SNMP agent and the iPlanet SNMP agent, a proxy
must be used that makes the Sun agent look like a subagent to the iPlanet
master agent.
193
Setting up SNMP Monitoring of the iPlanet Application
Server
Use your SNMP management tool to query the SNMP master agent on
port 161. You should see all the information provided by the Solaris SNMP
agent as well as any iPlanet subagents that you have configured.
The next time that you boot Solaris, the Sun and iPlanet SNMP agents will
be started automatically by the boot scripts that you have configured.
To set up the application server for SNMP monitoring:
1 Log in the the system.
Login to the system as root. (Only a root user can change the port number
and run the agents).
2 Change the port number for the Solaris SNMP agent:
a Stop the SNMP agent by running /etc/rc2.d/K76snmpdx stop.
b Edit /etc/rc3.d/S76snmpdx to run the Solaris daemon on a non-standard
port number, for example, 1161:
Replace
with
3 Configure and run the iPlanet agents, magt and sagt:
The master and proxy agents and startup scripts are found in <ias install
directory>\snmp.
a In the script S75snmpagt, add a line to the environment variable
GX_ROOTDIR so that it points to your iAS installation. For example, if
the iPlanet Application Server is installed in /usr/iplanet/ias6/ias:
/usr/lib/snmp/snmpdx -y -c /etc/snmp/conf
/usr/lib/snmp/snmpdx -p 1161 -y -c /etc/snmp/conf
GX_ROOTDIR=/usr/iplanet/ias6/ias
exprt GX_ROOTDIR
194
b Copy the script S75snmpagt to /etc/rc3.d
c chmod 755 /etc/rc3.d/S75snmpagt
d In /etc/rc3.d/S75snmpagt /etc/rc2.d/K07snmpagt
e You can configure system information and traps.
In the example below, information has been added about the system
owner and location, and SNMP traps have been sent to a network
manager station ("mde.uk.sun.com").
Note: There is no need to edit the proxy agent's configuration file
(CONFIG_SAGT).
f Start the iPlanet agents by running the command:
4 Start the Solaris SNMP agent:
Restart the Solaris SNMP agent by running the command:
5 Configure iPlanet Application Server for SNMP statistics:
a Start the iPlanet Application Server admin tool ksvradmin.
b In the General View, select the instance name that you want to manage.
c Click the SNMP tab in the management frame.
COMMUNITY public
ALLOW ALL OPERATIONS
INITIAL sysLocation "Under Joe Bloggs' Desk in Headquarters"
INITIAL sysContact "Joe Bloggs
Email: [email protected]
Voice: +1 650 555 1212"
MANAGER mde.uk.sun.com
SEND ALL TRAPS TO PORT 162
WITH COMMUNITY public
/etc/rc3.d/S75snmpagt start
/etc/rc3.d/S76snmpdx start
195
d Select Enable SNMP Administration and Monitoring and Enable SNMP
Debug.
e Type 60 in the Connection Attempt Interval field, and exit ksvradmin.
f Restart the iPlanet Application Server with the commands:
g Check in the logfile <iASInstallDir>/logs/ias.log that the application
server successfully connected to the master agent. You should see the
following line:
6 Start SNMP subagents for iPlanet Web Server:
a Use your Web browser to access the iPlanet Web Server.
b Select the Web server you wish to administer, and click the Manage
button.
c Select the Monitor tab, and click SNMP Subagent Configuration on the
left side of the page.
d Type in the configuration information and set the radio button Enable
SNMP Statistics Collection to "On".
e Click SNMP Subagent Control.
f Click the Start button.
7 Start SNMP subagents for iPlanet Directory Server:
a Use the Netscape Administration Console to manage the iPlanet
Directory Server.
b Select the Configuration tab.
c Click the SNMP tab in the Configuration frame.
d Select the Enable statistics collection check box.
e Set "Master Host" to "localhost".
iascontrol stop
iascontrol kill
iascontrol start
kas> SNMP: Connected to master agen
196
f Set "Master port" to 199.
g In the other fields, type the appropriate information.
h Click the Start Subagent button.
Configuring the iPlanet (NAS) Monitor
After you have configured the iPlanet SNMP Service, you must select the
counters that you want the iPlanet (NAS) monitor to measure. You select
these measurements using the iPlanet (NAS) Resources Choose
Measurements page.
To configure the iPlanet (NAS) Resources monitor:
Profile.
Click Monitors.
197
2 Under Web Application Server Resource Graphs, select iPlanet (NAS) and
3 Type the server name or IP address of the machine you want to monitor. If
the iPlanet SNMP agent is running on a different port than the default
SNMP port, you need to define the port number. Specify the port number
and the server name using the format:
For example,
display the sub-levels as a single string, separated by periods (.). In the
default view, the Choose Measurements page displays all nodes of the SNMP
objects in a tree-level hierarchy.
To improve the level of measurement information for the iPlanet (NAS)
<server name:port number>
digi:8888
198
For a list of the available performance counters, see iPlanet (NAS)
Note: The iPlanet (NAS) monitor can only monitor up to 25 measurements.
profile.
iPlanet (NAS) Performance Counters
Netscape Performance Counters
nasKesEngConnRetries The maximum number of times the administration
server will try to connect to an engine.
nasKesEngMaxRestart The maximum number of times the administration
server will restart an engine after a failure.
nasKesEngAutoStart Start all the engines at startup of the administration
server.
nasKesConfigHeartBeat Heart Beat.
199
KES Performance Counters
nasKesId The ID of the KES this engine belongs to.
nasKesMinThread The default minimum number of threads per
engine.
nasKesMaxThread The default maximum number of threads
per engine.
nasKesLoadBalancerDisable Enable or Disable the load balancer service.
nasKesCpuLoad The total CPU usage on this host.
nasKesDiskLoad The total disk usage on this host.
nasKesMemLoad The total memory usage on this host.
nasKesRequestLoad The number of requests on this NAS.
nasKesCpuLoad
Factor
The relative importance of CPU usage in
computing the server load. This number is
specified as a percentage. The sum of all
server load factors, CPULoad, DiskLoad,
MemLoad and ExecReqs must equal 100%.
nasKesDiskLoadFactor The relative importance of Disk usage in
nasKesMemLoadFactor The relative importance of Memory usage in
200
nasKesAppLogicsRunningFactor The relative importance of the number of
times an AppLogic is run in computing the
AppLogic execution performance. This figure
is specified as a percentage. The sum of all
agent load factors, ResultCached,
AvgExecTime, LastExecTime, and ServerLoad
must equal 100%.
nasKesResultsCachedFactor The relative importance of the cached results
of an AppLogic in computing the AppLogic
execution performance. This figure is
must equal 100%
nasKesAvgExecTimeFactor The relative importance of the average
execution time of an AppLogic in computing
the AppLogic execution performance. This
figure is specified as a percentage. The sum
of all agent load factors, ResultCached,
must equal 100%.
nasKesLastExecTimeFactor The relative importance of the last execution
time of an AppLogic in computing the
must equal 100%.
nasKesHitsFactor The relative importance of the number of
AppLogics running in computing the
must equal 100%.
201
nasKesServerLoadFactor The relative importance of the server load
(computed using the four server load factors)
in computing AppLogic execution
performance. The sum of all agent load
factors, ResultCached, AvgExecTime,
LastExecTime, and ServerLoad must equal
100%.
nasKesBroadcastInterval The length of time in seconds, between each
broadcast attempt from the load balancer
daemon.
nasKesApplogicBroadcastInterval The length of time in seconds, between each
broadcast of AppLogics load information
across all the server in the cluster. This
should be greater than
nasKesBroacastInterval.
nasKesServerBroadcastInterval The length of time in seconds, between each
broadcast of server load information across
all the server in the cluster. This should be
greater than nasKesBroacastInterval.
nasKesServerLoadUpdateInterval The length of time in seconds between each
update of server load informations. A server
load update applies the server load data that
has been sampled up until the moment
when the update occurs.
nasKesCpuLoadUpdateInterval The length of time, in seconds, between
each sampling of CPU usage.
nasKesDiskLoadUpdateInterval The length of time, in seconds, between
each sampling of disk usage.
nasKesMemLoadUpdateInterval The length of time, in seconds, between
each sampling of memory thrashes.
nasKesTotalReqsUpdateInterval The length of time, in seconds, between
each sampling of the number of requests.
nasKesMaxHops The maximum number of times a request
can be load-balanced to another server.
202
nasKesODBCReqMinThread The minimum number of threads reserved
to process asynchronous requests.
nasKesODBCReqMaxThread The maximum number of threads reserved
nasKesODBCCacheMaxConns The maximum number of connections
opened between NAS and the database.
nasKesODBCCacheFreeSlots The minimum number of cached
connections established between NAS and
the database.
nasKesODBCCacheTimeout The time after which an idle connection is
dropped.
nasKesODBCCacheInterval The interval in seconds at which the cache
cleaner will try to disconnect connections
already idle for longer than the specified
timeout.
nasKesODBCConnGiveupTime Maximum time the driver will try to connect
to the database.
nasKesODBCCacheDebug Turns on the connection cache debug
information.
nasKesODBCResultSetInitRows The number of rows fetched at once from
the database.
nasKesODBCResultSetMaxRows The maximum number of rows the cached
result set can contain.
nasKesODBCResultSetMaxSize The maximum size of result set the driver
will cache.
nasKesODBCSqlDebug Turns on SQL debug information.
nasKesODBCEnableParser Turns on SQL parsing.
nasKesORCLReqMinThread The minimum number of threads reserved
nasKesORCLReqMaxThread The maximum number of threads reserved
203
nasKesORCLCacheMaxConns The maximum number of connections
nasKesORCLCacheFreeSlots The minimum number of cached
the database.
nasKesORCLCacheTimeout The time after which an idle connection is
dropped.
nasKesORCLCacheInterval The interval in seconds at which the cache
cleaner will try to disconnect connections
already idle for longer than the specified
timeout.
nasKesORCLConnGiveupTime The maximum time the driver will spend
trying to obtain a connection to Oracle.
nasKesORCLCacheDebug Turns on the connection cache debug
information.
nasKesORCLResultSetInitRows The number of rows fetched at once from
the database.
nasKesORCLResultSetMaxRows The maximum number of rows the cached
nasKesORCLResultSetMaxSize The maximum size of result set the driver
will cache.
nasKesORCLSqlDebug Turns on SQL debug information.
nasKesSYBReqMinThread The minimum number of threads reserved
nasKesSYBReqMaxThread The maximum number of threads reserved
to process asynchronous request.
nasKesSYBCacheMaxConns The maximum number of connections
nasKesSYBCacheFreeSlots The minimum number of cached
the database.
204
Engine Performance Counters
nasKesSYBCacheTimeout The time after which an idle connection is
dropped.
nasKesSYBCacheInterval The interval time between cached
connections.
nasKesSYBConnGiveupTime The maximum time the driver will spend
trying to obtain a connection to Sybase
before giving up.
nasKesSYBCacheDebug Turns on the connection cache debug
information.
nasKesSYBResultSetInitRows The number of rows fetched at once from
the database.
nasKesSYBResultSetMaxRows The maximum number of rows the cached
nasKesSYBResultSetMaxSize The maximum size of result set the driver
will cache.
nasEngKesPort The port of the KXS this engine serves. This
is supplied as part of the object ID and
cannot be modified after creation.
nasEngPort The TCP/IP port this engine is listening on.
The port can only be specified at the
creation of the engine. It is not allowed to
modify it.
nasEngType Type of the engine: executive(0), Java(1000),
C++(3000).
nasEngId The ID is an incremental number starting at
0. The ID cannot be modified.
205
nasEngName The name of this engine. This is an
informational string that contains kcs, kxs,
or kjs.
nasEngNewConsole Starts each engine in a new console window.
nasEngStatus The status column used to add, remove,
enable or disable an engine. To create an
engine, one needs to set. This follows
rfc1443.
nasEngMinThread The default minimum number of threads per
engine.
nasEngMaxThread The default maximum number of threads
per engine.
nasEngReqRate The rate at which requests arrive.
nasEngTotalReq The total number of requests processed since
engine startup.
nasEngReqNow The number of requests being processed.
nasEngReqWait The requests waiting to be serviced.
nasEngReqReady The requests that are ready to be serviced.
nasEngAvgReqTime The average request processing time.
nasEngThreadNow Number of threads in use by the request
manager.
nasEngThreadWait The number of idle threads.
nasEngWebReq
Queue
The number of web requests that are queued.
nasEngFailedReq The number of requests that failed.
nasEngTotalConn The total number of connections opened.
nasEngTotalConnNow The total number of connections in use.
nasEngTotalAccept The total number of connections listening to
incoming requests.
206
nasEngTotalAcceptNow The total number of connections listening to
incoming connections in use.
nasEngTotalSent The total number of packets sent.
nasEngTotalSentBytes The total number of bytes sent.
nasEngTotalRecv The total number of packets received.
nasEngTotalRecvBytes The total number of bytes received.
nasEngBindTotal The number of AppLogic bound since
startup.
nasEngBindTotalCached The number of AppLogic cached since
startup.
nasEngTotalThreads Total number of threads created in this
process.
nasEngCurrentThreads Total number of threads in use in this
process.
nasEngSleepingThreads Number of threads sleeping in this process.
nasEngDAETotalQuery Total number of queries executed since
startup.
nasEngDAEQueryNow The number of queries being processed.
nasEngDAETotalConn The number of logical connections created
since startup.
nasEngDAEConnNow The number of logical connections in use.
nasEngDAECacheCount The number of caches.
nasEngODBCQueryTotal Total number of queries executed since
startup.
nasEngODBCPreparedQueryTotal Total number of odbc prepared queries
executed since startup.
nasEngODBCConnTotal Total number of connections opened since
startup.
nasEngODBCConnNow Number of connections currently opened.
207
nasEngORCLQueryTotal Total number of queries executed since
startup.
nasEngORCLPreparedQueryTotal Total number of prepared queries executed
since startup.
nasEngORCLConnTotal Total number of connections established
with Oracle since startup.
nasEngORCLConnNow Number of connections opened with Oracle
now.
nasEngSYBQueryTotal Total number of queries the driver processed
since startup.
nasEngSYBPreparedQueryTotal Total number of prepared queries processed
since startup.
nasEngSYBConnTotal Total number of connections opened since
startup.
nasEngSYBConnNow Number of SYB connections opened now.
nasStatusTrapEntry The KES definition.
nasTrapKesIpAddress The IP Address of KES host.
nasTrapKesPort The port of the main engine of this NAS.
nasTrapEngPort The port of the engine generating this event.
nasTrapEngState The port of the engine generating this event.
208
209
26
Microsoft Active Server Pages Monitoring
The Microsoft Active Server Pages (ASP) monitor displays statistics about the
resource usage on the ASP server during the load test run.
Configuring the Microsoft Active Server Pages Monitor on page 209
MS Active Server Pages Performance Counters on page 212
Configuring the Microsoft Active Server Pages Monitor
You select the measurements you want to monitor on the Microsoft ASP
application server using the MS Active Server Pages Choose Measurements
page.
Note: To monitor an ASP server through a firewall, use TCP, port 139.
To configure the ASP monitor:
Profile.
Chapter 26 Microsoft Active Server Pages Monitoring
210
Click Monitors.
2 Under Web Application Server Resource Graphs, select MS Active Server
Pages and click Next. The Choose Server page opens.
monitor.
211
For a list of the available performance counters, see MS Active Server Pages
profile.
212
MS Active Server Pages Performance Counters
Errors per Second The number of errors per second.
Requests Wait Time The number of milliseconds the most recent request
was waiting in the queue.
Requests Executing The number of requests currently executing.
Requests Queued The number of requests waiting in the queue for
service.
Requests Rejected The total number of requests not executed because
there were insufficient resources to process them.
Requests Not Found The number of requests for files that were not
found.
Requests/sec The number of requests executed per second.
Memory Allocated The total amount of memory, in bytes, currently
allocated by Active Server Pages.
Errors During Script
Run-Time
The number of failed requests due to run-time
errors.
Sessions Current The current number of sessions being serviced.
Transactions/sec The number of transactions started per second.
213
27
Oracle9iAS HTTP Server Monitoring
The Oracle9iAS HTTP Server monitor displays statistics about the resource
usage on the Oracle9iAS HTTP server during the load test run.
Configuring the Oracle9iAS HTTP Server Monitor on page 213
Oracle9iAS HTTP Server Performance Counters on page 217
Configuring the Oracle9iAS HTTP Server Monitor
The Oracle9iAS HTTP Server (SiteScope) monitor allows you to monitor
multiple parameters or counters with a single monitor instance. This allows
you to watch server loading for performance, availability, and capacity
planning.
Before Running the Load Test
You need to configure the Oracle9iAS HTTP Server online monitor, and
select the default measurements you want to display. You select
measurements to monitor the Oracle9iAS HTTP server using the Oracle
HTTP Server Monitor Configuration dialog box.
For Oracle9iAS HTTP Server monitor:
You must start running the Oracle9iAS HTTP server before you begin
selecting the measurements you want to monitor.
The port you use to monitor an Oracle9iAS HTTP server through a firewall
Chapter 27 Oracle9iAS HTTP Server Monitoring
214
For Oracle9iAS HTTP Server (SiteScope) monitor:
Create a separate monitor instance for each Oracle9i application server in
your environment.
You must enable Web caching on the Oracle9i application server in order to
use the Oracle9iAS monitor.
The default run schedule for this monitor is every 10 minutes, but you can
change it to run more or less often using the Update every setting.
Although SiteScope is installed on the Utility Server, we recommend
installing it on the same machine as the Controller, or on a dedicated server
in order to reduce load on the Utility Server.
To configure the Oracle9iAS HTTP monitor:
Profile.
Click Monitors.
2 Under Web Application Server Resource Graphs, select Oracle9iAS HTTP
Server or Oracle9iAS HTTP Server (SiteScope) and click Next. The Choose
Server page opens.
215
For Oracle9iAS HTTP Server, type the user login name and password on
the Oracle server.
For Oracle9iAS HTTP Server (SiteScope), type the SiteScope machine
name and port (default: 8888), and specify if you are using a Secure HTTP
Click Next.
4 (Oracle9iAS HTTP Server (SiteScope) monitor only) The first time you add
measurements using the Oracle9iAS HTTP Server (SiteScope) monitor you
need to configure the monitor properties.
On the Configuring Oracle9iAS Monitor page that opens, enter the monitor
properties, as follows:
Port. Type the port number of the Oracle9iAS server.
URL. The server administration URL appears by default, and has the
following form:
http://server:port/webcacheadmin?SCREEN_ID=CGA.Site.Stats&ACTION=
Show
HTTP Proxy. If you are using a proxy to access the Oracle9iAS server, type
the domain name and port of an HTTP Proxy Server (for example,
proxy.sitescope.com:8080).
Authorization User Name. If the URL requires authentication, type the
user name for the server administrator page.
Authorization Password. If the URL requires authentication, type the
password for the server administrator page.
Proxy Server User Name. If the proxy server requires authorization, type
the user name. Your proxy server must support Proxy-Authenticate for
these options to function.
Update every. Type how frequently (in seconds) the monitor should
check the Oracle9iAS server. You must specify a time increment of at least
15 seconds. The default is 30 seconds.
216
Click Next.
For a list of the available performance counters, see Oracle9iAS HTTP Server
profile.
217
Oracle9iAS HTTP Server Performance Counters
The following tables describe the performance counters for the Oracle9i
Server monitors.
Oracle9iAS HTTP Server (SiteScope) Monitor Performance Counters
Performance Center Oracle9iAS HTTP Server Monitor Performance Counters
Oracle9iAS HTTP Server (SiteScope) Monitor Performance
Counters
Note: The Oracle9iAS HTTP Server (SiteScope) monitor can monitor a
maximum of 10 counters at one time.
The following table describes the counters that can be monitored by the
Oracle9iAS HTTP Server (SiteScope) monitor:
Up/Down Time (up/down) Status of the application Web server.
Completed Requests
(number/sec)
Number of requests that the application Web
server processes per second.
Completed
Requests(max/sec)
Maximum number of requests that the
application Web server can process per second.
Completed Requests
(avg/sec)
Average number of requests that the application
Web server has processed per second.
Completed Requests (total) Accumulated number of requests that the
application Web server has processed.
Latency (avg this interval) Average latency for 10 second intervals to
process requests for Oracle Web Cache.
Latency (avg since start) Average number of seconds to process requests
for Oracle Web Cache since the application Web
server started.
218
Load (now) Current number of connections from Oracle
Web Cache that the application Web server has
open.
Load (max) Maximum number of connections that the
application Web server has had open at one
time.
Active Sessions (now) Current number of active connections from
Oracle Web Cache to the application Web
servers.
Active Sessions (max) Maximum number of active connections that
the application Web server has had open at one
time.
Apology Pages Served
(Network Error - number this
second)
Current number of apology pages that the
Oracle Web Cache is serving to Web browsers
due to a network error.
Apology Pages Served
(Network Error - Total)
Total number of apology pages that the Oracle
Web Cache is serving to Web browsers due to a
network error.
Apology Pages Served (Site
Busy - number this second)
Current number of apology pages that the
Oracle Web Cache is serving to Web browsers
due to a busy Web site error.
Apology Pages Served (Site
Busy - total)
Total number of apology pages that the Oracle
Web Cache is serving to Web browsers due to a
busy Web site error.
Application Web Server
Backlog (now)
Current number of requests that the application
Web server is processing for Oracle Web Cache.
Application Web Server
Backlog (max)
Maximum number of requests that the
application Web server has processed for Oracle
Web Cache.
219
Performance Center Oracle9iAS HTTP Server Monitor
Performance Counters
The following tables describe some of the modules and the counters that
can be monitored by the native Performance Center Oracle9iAS HTTP Server
monitor.
Modules
mod_mime.c Determines document types using file extensions
mod_mime_magic.c Determines document types using "magic numbers"
mod_auth_anon.c Provides anonymous user access to authenticated areas
mod_auth_dbm.c Provides user authentication using DBM files
mod_auth_digest.c Provides MD5 authentication
mod_cern_meta.c Supports HTTP header metafiles
mod_digest.c Provides MD5 authentication (deprecated by
mod_auth_digest)
mod_expires.c Applies Expires: headers to resources
mod_headers.c Adds arbitrary HTTP headers to resources
mod_proxy.c Provides caching proxy abilities
mod_rewrite.c Provides powerful URI-to-filename mapping using
regular expressions
mod_speling.c Automatically corrects minor typos in URLs
mod_info.c Provides server configuration information
mod_status.c Displays server status
mod_usertrack.c Provides user tracking using cookies
mod_dms.c Provides access to DMS Apache statistics
mod_perl.c Allows execution of perl scripts
mod_fastcgi.c Supports CGI access to long-lived programs
220
mod_ssl.c Provides SSL support
mod_plsql.c Handles requests for Oracle stored procedures
mod_isapi.c Provides Windows ISAPI extension support
mod_setenvif.c Sets environment variables based on client
information
mod_actions.c Executes CGI scripts based on media type or request
method
mod_imap.c Handles imagemap files
mod_asis.c Sends files that contain their own HTTP headers
mod_log_config.c Provides user-configurable logging replacement for
mod_log_common
mod_env.c Passes environments to CGI scripts
mod_alias.c Maps different parts of the host file system in the
document tree, and redirects URLs
mod_userdir.c Handles user home directories
mod_cgi.c Invokes CGI scripts
mod_dir.c Handles the basic directory
mod_autoindex.c Provides automatic directory listings
mod_include.c Provides server-parsed documents
mod_negotiation.c Handles content negotiation
mod_auth.c Provides user authentication using text files
mod_access.c Provides access control based on the client hostname
or IP address
mod_so.c Supports loading modules (.so on UNIX, .dll on
Win32) at run time
mod_oprocmgr.c Monitors JServ processes and restarts them if they fail
221
Counters
mod_jserv.c Routes HTTP requests to JServ server processes.
Balances load across multiple JServs by distributing
new requests in round-robin order
mod_ose.c Routes requests to the JVM embedded in Oracles
database server
http_core.c Handles requests for static Web pages
handle.minTime The minimum time spent in the module handler
handle.avg The average time spent in the module handler
handle.active The number of threads currently in the handle
processing phase
handle.time The total amount of time spent in the module
handler
handle.completed The number of times the handle processing phase
was completed
request.maxTime The maximum amount of time required to service
an HTTP request
request.minTime The minimum amount of time required to service
an HTTP request
request.avg The average amount of time required to service an
HTTP request
request.active The number of threads currently in the request
processing phase
request.time The total amount of time required to service an
HTTP request
request.completed The number of times the request processing phase
was completed
222
connection.maxTime The maximum amount of time spent servicing any
HTTP connection
connection.minTime The minimum amount of time spent servicing any
HTTP connection
connection.avg The average amount of time spent servicing HTTP
connections
connection.active The number of connections with currently open
threads
connection.time The total amount of time spent servicing HTTP
connections
connection.completed The number of times the connection processing
phase was completed
numMods.value The number of loaded modules
childFinish.count The number of times the Apache parent server
started a child server, for any reason
childStart.count The number of times "children" finished
"gracefully." There are some ungraceful error/crash
cases that are not counted in childFinish.count
Decline.count The number of times each module declined HTTP
requests
internalRedirect.count The number of times that any module passed
control to another module using an "internal
redirect"
cpuTime.value The total CPU time utilized by all processes on the
Apache server (measured in CPU milliseconds)
heapSize.value The total heap memory utilized by all processes on
the Apache server (measured in kilobytes)
pid.value The process identifier of the parent Apache process
upTime.value The amount of time the server been running
(measured in milliseconds)
223
28
SilverStream Monitoring
The SilverStream monitor displays statistics about the resource usage on the
SilverStream server during the load test run.
Configuring the SilverStream Monitor on page 223
SilverStream Performance Counters on page 226
Configuring the SilverStream Monitor
To monitor a SilverStream server you need to know the server statistics
information URL. A simple way to verify the statistics URL is to access it
from a browser.
The URL should be in the format:
For example:
To configure the SilverStream monitor:
http://<server_name/IP_address>:<port_number>/SilverStream/Statistics
http://199.203.78.57:80/SilverStream/Statistics
Chapter 28 SilverStream Monitoring
224
Profile.
Click Monitors.
2 Under Web Application Server Resource Graphs, select SilverStream and
3 Type the details of the server whose resources you want to monitor
a Type the name or IP address.
b In the URL box, type the URL or accept the default URL:
c In the Port box, type the Port number (without the server name), or
accept the default port.
(/SilverStream/Statistics)
225
another. Please consult the Web server administrator.
For a list of the available performance counters, see SilverStream
profile.
Note: To monitor a SilverStream server through a firewall, use the Web
server port (by default, port 80).
226
SilverStream Performance Counters
can be monitored:
#Idle Sessions The number of sessions in the Idle state.
Avg. Request processing
time
The average request processing time.
Bytes Sent/sec The rate at which data bytes are sent from the Web
server.
Current load on Web
Server
The percentage of load utilized by the SilverStream
server, scaled at a factor of 25.
Hits/sec The HTTP request rate.
Total sessions The total number of sessions.
Free memory The total amount of memory in the Java Virtual
Machine currently available for future allocated
objects.
Total memory The total amount of memory in the Java Virtual
Machine.
Memory Garbage
Collection Count
The total number of times the JAVA Garbage
Collector has run since the server was started.
Free threads The current number of threads not associated with a
client connection and available for immediate use.
Idle threads The number of threads associated with a client
connection, but not currently handling a user
request.
Total threads The total number of client threads allocated.
227
29
WebLogic (JMX) Monitoring
The BEA WebLogic (JMX) monitor uses the Java JMX interface to access
runtime MBeans on the server. An MBean is a container that holds the
performance data.
Configuring the WebLogic (JMX) Monitor on page 231
BEA WebLogic (JMX) Performance Counters on page 235
Before using the WebLogic (JMX) monitor, you must:
Install Java 1.3 or later on the Performance Center Host machine. You must
install Java 1.3 or later to work with Weblogic 6.x or 7.x, and Java 1.4 to
work with WebLogic 8.1.
Set Permissions for user to be able to monitor MBeans
Make sure that classes load directly from the file system if the servlet is
disabled or not used. For more information, see Loading Classes from the
Server on page 230.
Java 1.3.x:
1 Install Java 1.3 or later on the Performance Center Host machine. If Java 1.3
or later is already installed, but is not the default Java version being used,
specify the full path to the updated version.
Chapter 29 WebLogic (JMX) Monitoring
228
2 Specify the path in the <Performance Center root
folder>\dat\monitors\WebLogicMon.ini file. Edit the JVM entry in the
[WebLogicMon] section. For example:
Java 1.4:
1 Install JDK 1.4 on the Performance Center Host machine. If Java 1.4 is
already installed, but is not the default Java version being used, specify the
full path to the updated version in the command line.
2 Copy weblogic.jar from the <WebLogic Server installation>\..\lib to
<Performance Center root folder>\classes.
3 Remove jmxri.jar from <Performance Center root folder>\classes.
4 Specify the path in the <Performance Center root
folder>\dat\monitors\WebLogicMon.ini file. Edit the JVM entry in the
[WebLogicMon] section. For example:
5 Edit the JavaVersion entry in the [WebLogicMon] section.
Note: To use the WebLogic (JMX) monitor, you must make sure that
WebLogic 6.0 or above is installed on your server.
Setting Permissions for Monitoring
You must set certain permissions for a user to be able to monitor MBeans.
To log on to the WebLogic (JMX) server:
Type the user name and password of a user with administrative privileges.
To set permissions for WebLogic version 6.x:
1 Open the WebLogic console (http://<host:port>/console).
2 In the tree on the left, select Security > ACLs.
JVM="E:\Program Files\JavaSoft\JRE\1.3.1\bin\javaw.exe
JVM="E:\Program Files\JavaSoft\JRE\1.4\bin\javaw.exe
229
If you are working with the WebLogic 6.1 console, click Create a new ACL...
in the area on the right.
3 In the New ACL Name box, type weblogic.admin.mbean, and click Create.
If you are working with the WebLogic 6.1 console, click Add a new
Permission... in the area on the right.
4 In the New Permission box (or Permission box, in the WebLogic 6.1
console), type access. In the WebLogic 6.0 console, click Create.
5 In the Users box and Groups box, type the name of any user or group you
want to use for monitoring.
6 Click Grant Permission in the WebLogic 6.0 console. In the WebLogic 6.1
console, click Apply.
To set permissions for WebLogic versions 7.x and 8.x:
1 Open the WebLogic console (http://<host:port>/console).
2 In the tree on the left, select Security > Realms > myrealm > Users, and click
Configure a new User... in the area on the right. The Create User: General
tab opens.
3 In the Name box, type weblogic.admin.mbean, type a password, confirm the
password, then click Apply.
4 In the Groups tab, type the name of any user or group you want to use for
monitoring, then click Apply.
230
Loading Classes from the Server
The WebLogic (JMX) monitor utilizes a built-in server called the
ClasspathServlet to load classes directly and automatically from the server.
The advantages of this are easy installation and version independence. The
disadvantages are a slight decrease in performance when loading classes for
the first time (due to the size of the servlet), and the possibility of the servlet
becoming disabled.
If the servlet is disabled, or if you do not want to use the servlet, you can
load classes directly from the file system.
Note: For WebLogic version 8.1, you must load classes directly from the file
system, as the WebLogic (JMX) monitor will not load the classes
automatically.
To load classes directly from the file system:
1 Copy the weblogic.jar file from the application server install folder (under
the lib folder) to <Performance Center root folder>\classes.
2 If the classes file is not located in the default <Performance Center root
folder> folder, you need to specify the full path to it in the <Performance
Center root folder>\dat\monitors\WebLogicMon.ini file. In this file,
change the line
to
Weblogic=weblogic.jar
Weblogic=<full path to weblogic.jar>
231
Configuring the WebLogic (JMX) Monitor
You select the measurements you want to monitor on the WebLogic (JMX)
application server using the WebLogic JMX Choose Measurements page.
Note: The WebLogic (JMX) monitor utilizes a built-in server called the
ClasspathServlet to load classes directly and automatically from the server.
For information about loading classes from the server, see Loading Classes
from the Server on page 230.
To configure the WebLogic (JMX) Monitor:
Profile.
Click Monitors.
2 Under Web Application Server Resource Graphs, select WebLogic (JMX) or
WebLogic (JMX) (SiteScope) and click Next. The Choose Server page opens.
232
3 Type the server name or IP address of the machine you want to monitor
For example,
In addition:
For WebLogic (JMX), type the user login name and password on the
WebLogic JMX machine.
For WebLogic (JMX) (SiteScope), type the SiteScope machine name and
port (default: 8888), and specify if you are using a Secure HTTP
Click Next.
4 (WebLogic (JMX) (SiteScope) monitor only) The first time you add
measurements using the WebLogic (JMX) (SiteScope) monitor you need to
configure the monitor properties.
In the Configuring SNMP by MIB Monitor page that opens, enter the
monitor properties ,as follows:
Port Number. Type the port number of WebLogic (JMX) server. The
default port is 7001.
Password. Type the password required to log on to the WebLogic server.
User Name. Type the user name required to log on to the WebLogic
server.
Secure Server. Check this box if you are using a secure server connection
option. If you select this option, you must type the applicable port
number used by the WebLogic server for secure connections. The default
secure server port is 7002.
Server. Type the name of the server where the WebLogic JMX application
is running.
mercury:8111
233
WebLogic Jar File. Type the absolute path name to the weblogic.jar file
on the SiteScope machine. This file must be installed on the SiteScope
server and can be downloaded from the WebLogic server. An example
path is:
This file is not strictly required for monitoring some earlier versions of
WebLogic 6. In this case, leaving this box blank normally causes any
necessary classes to be downloaded directly from the WebLogic server.
This is not as efficient as loading the classes from the a *.jar file on the
server where SiteScope is running.
Note: Do not install the weblogic.jar file in the SiteScope directory tree.
For example, do not install it in the <SiteScope install
path>/SiteScope/java/lib/ext directory as this will cause the Weblogic
monitor to fail. You must create a separate directory for this file on the
server where SiteScope is running.
WLCipher Jar File. (This option is used only with the Secure Server (SSL)
option.)
For some versions of the WebLogic Server, you need to type the absolute
path to the wlcipher.jar file on the SiteScope machine.
For example:
WebLogic License File. (This option is used only with the Secure Server
(SSL) option.)
Type the absolute path to a BEA license file on the SiteScope machine.
For example:
c:\bea\weblogic7\ebcc\lib\ext\weblogic.jar
c:bea\weblogic81\lib\wlciher.jar
c:\bea\license.bea
234
Location of JVM. Specify the location of a JVM where the WebLogic
monitoring process should run.
Note: For monitors which do not use the Secure Server (SSL) option, this
is not required. For monitors that do use SSL, the version of the JVM
must be 1.4.1 or earlier.
Timeout. Type the total time, in seconds, that SiteScope should wait for
all WebLogic requests (including retries) to complete. The default is 180
seconds.
Update every. Type how frequently the monitor should read the server
statistics. The list to the right of the text box lets you select time
increments of seconds, minutes, hours, or days. You must specify a time
increment of at least 15 seconds. The default is 30 seconds.
Click Next.
For a list of the available performance counters, see BEA WebLogic (JMX)
profile.
235
BEA WebLogic (JMX) Performance Counters
The following measurements are available for the WebLogic (JMX) server:
LogBroadcasterRuntime
ServerRuntime
For more information on the measurements contained in each of the
following measurement categories, see HPs Load Testing Monitors Web site
(http://www.mercury.com/us/products/performance-
center/loadrunner/monitors/bealogic.html).
ServletRuntime
WebAppComponentRuntime
EJBStatefulHomeRuntime
JTARuntime
JVMRuntime
EJBEntityHomeRuntime.
DomainRuntime
EJBComponentRuntime
DomainLogHandlerRuntime
JDBCConnectionPoolRuntime
ExecuteQueueRuntime
ClusterRuntime
MessagesLogged The number of total log messages generated
by this instance of the WebLogic server.
Registered Returns "false" if the MBean represented by
this object has been unregistered.
CachingDisabled Private property that disables caching in
proxies.
236
JMSRuntime
TimeServiceRuntime
EJBStatelessHomeRuntime
WLECConnectionServiceRuntime
ServerSecurityRuntime
UnlockedUsersTotalCount Returns the number of times a user
has been unlocked on the server
InvalidLoginUsersHighCount Returns the high-water number of
users with outstanding invalid login
attempts for the server
LoginAttemptsWhileLockedTotalCount Returns the cumulative number of
invalid logins attempted on the
server while the user was locked
Registered Returns "false" if the MBean
represented by this object has been
unregistered
LockedUsersCurrentCount Returns the number of currently
locked users on the server
CachingDisabled Private property that disables
caching in proxies
InvalidLoginAttemptsTotalCount Returns the cumulative number of
invalid logins attempted on the
server
UserLockoutTotalCount Returns the cumulative number of
user lockouts done on the server
237
30
WebSphere Monitoring
The WebSphere monitor displays statistics about the resource usage on the
WebSphere 3.x, 4.x, or 5.x servers during the load test run.
Configuring the WebSphere Monitor on page 238
WebSphere Performance Counters on page 243
The WebSphere 3.x, 4.x, and 5.x application servers have different monitor
installation requirements.
To monitor WebSphere version 5.x, you need to deploy the performance
servlet on the application server using the IBM WebSphere "Installing a New
Application" wizard.
To deploy the performance servlet on the application server for
WebSphere 5.x:
1 From the administrative console, click Applications > Install New
Application in the console navigation tree.
2 For Path, specify the full path name of the source application file
("PerfServletApp.ear") on the server machine and click Next.
3 Select the Generate Default Bindings check box and click Next.
4 On the Install New Application page, click Summary, and select the
Cell/Node/Server option. Click Click here.
Chapter 30 WebSphere Monitoring
238
5 On the Map modules to application servers panel, select the server onto
which the application files will install from the Clusters and Servers list, and
select Module to select all of the application modules.
6 Click Next, and in the Summary panel click Finish.
7 Verify that the servlet is running properly and that the performance data is
being generated. A simple way to verify that the performance data is
accessible is to display it in a Web browser. The URL should be in the format:
For example:
Note: Only browsers that are XML-compatible will allow you to view the
performance XML file.
Configuring the WebSphere Monitor
You select measurements to monitor the WebLogic application server using
the WebLogic Choose Measurements page.
If you are using the SiteScope monitor engine, SiteScope must be installed
on a server. We recommend installing it on the same machine as the
Controller, or on a dedicated server in order to reduce load on the Utility
Server.
To configure the WebSphere or WebSphere 4.x-5.x monitor:
http://<server name:port number>/<servlet_folder>/com.ibm.ivb.epm.servlet.
PerformanceServlet
http://websphere.mercury.co.il:81/servlet/com.ibm.ivb.epm.servlet.
PerformanceServlet
239
Profile.
Click Monitors.
2 Under Web Application Server Resource Graphs, select WebSphere or
WebSphere 4.x-5.x and click Next. The Choose Server page opens.
3 Type the details of the server whose resources you want to monitor.
a Type the name or IP address.
b Type the user login name and password.
240
For a list of the available performance counters, see WebSphere
profile.
Note: The port you use to monitor a WebSphere server through a firewall
241
Specifying a Web Alias for the Servlet Directory
You can monitor as many application servers as you want by specifying a
Web alias for the servlet directory.
To specify another Web alias for the servlet directory:
By default, Performance Center uses the alias servlet as the servlet directory
Web alias. For example, if the WebSphere Server machine is named mercury
and the path for the servlets directory is
E:\AppServer\hosts\default_host\default_app\servlets, Performance Center will
request the XML file in the URL:
where servlet is the default web alias for the servlet directory.
If the Web alias for the servlet directory is not servlet, you must specify the
servlet directory Web alias in the Add Machine dialog box according to the
format:
For example,
Using this method, you can monitor as many application servers as you
wantwhether they are installed on the same machine, or on different
machines.
To monitor other applications, in addition to the default application:
You can monitor as many applications as you want, regardless of whether
they are installed on the same machine or different machines.
1 Copy the same files that you copied to the Servlets directory for the Default
application to the Servlets directories for any other Web applications that
you want to monitor.
2 Add the com.ibm.ivb.epm.servlet.PerformanceServlet to the configuration
in the WebSphere Console for each Web application.
http://mercury/servlet/com.ibm.ivb.epm.servlet.PerformanceServlet
http://<server name:port number>/<servlet_dir_alias>
http://mercury/servlet2
242
3 Add the Web application to be monitored to the WebSphere Performance
Monitor using the format:
For example,
Enabling the EPM Counters on the WebSphere 3.5.x
Server
To work with WebSphere version 3.5.x
1 The EPM counters in 3.5.x are by default set to none. To enable the counters,
select the application server you are monitoring in the WebSphere
Administrators Console browser.
2 Right-click the application server and select Performance. Select
Performance Modules from the pop-up window.
3 Right-click Performance Modules to select a performance level. Selecting
various levels of counters enables the application server to manage varying
levels of performance data.
4 Click the Set button.
5 In versions 3.5.2 and 3.5.3 the Servlet counters have been disabled. To
enable the Servlet counters, you need to modify the contents of the
com/ibm/servlet/appserver.properties file located in
"<WAS_HOME>\lib\ibmwebas.jar".
Extract the jar file and modify the appserver.properties as follows:
http://<server:port_number>/<servlet_dir_alias>/servlet
http://mercury/servlet3/servlet
#listeners.application=com.ibm.servlet.engine.EPMApplicationListener
com.ibm.servlet.debug.OLTServletManager
listeners.application=
243
Should be:
6 Repackage the jar file.
WebSphere Performance Counters
Run-Time Resources
Contains resources related to the Java Virtual Machine run-time, as well as
the ORB.
BeanData
Every home on the server provides performance data, depending upon the
type of bean deployed in the home. The top level bean data holds an
aggregate of all the containers.
listeners.application=com.ibm.servlet.engine.EPMApplicationListener
com.ibm.servlet.debug.OLTServletManager
#listeners.application=
MemoryFree The amount of free memory remaining in the Java
Virtual Machine
MemoryTotal The total memory allocated for the Java Virtual
Machine
MemoryUse The total memory in use within the Java Virtual
Machine
BeanCreates The number of beans created. Applies to an
individual bean that is either stateful or entity
EntityBeanCreates The number of entity beans created
244
BeanRemoves The number of entity beans pertaining to a
specific bean that have been removed. Applies to
an individual bean that is either stateful or
entity
EntityBeanRemoves The number of entity beans removed
StatefulBeanCreates The number of stateful beans created
StatefulBeanRemoves The number of stateful bean removed
BeanPassivates The number of bean passivates pertaining to a
specific bean. Applies to an individual bean that is
either stateful or entity
EntityBeanPassivates The number of entity bean passivates
StatefulBeanPassivates The number of stateful bean passivates
BeanActivates The number of bean activates pertaining to a
specific bean. Applies to an individual bean that is
either stateful or entity
EntityBeanActivates The number of entity bean activates
StatefulBeanActivates The number of stateful bean activates
BeanLoads The number of times the bean data was loaded.
Applies to entity
BeanStores The number of times the bean data was stored in
the database. Applies to entity
BeanInstantiates The number of times a bean object was created.
This applies to an individual bean, regardless of
its type.
StatelessBeanInstantiates The number of times a stateless session bean
object was created
StatefulBeanInstantiates The number of times a stateful session bean object
was created
EntityBeanInstantiates The number of times an entity bean object was
created
245
BeanDestroys The number of times an individual bean object
was destroyed. This applies to any bean,
regardless of its type
StatelessBeanDestroys The number of times a stateless session bean
object was destroyed
StatefulBeanDestroys The number of times a stateful session bean object
was destroyed
EntityBeanDestroys The number of times an entity bean object was
destroyed
BeansActive The average number of instances of active beans
pertaining to a specific bean. Applies to an
individual bean that is either stateful or entity
EntityBeansActive The average number of active entity beans
StatefulBeansActive The average number of active session beans
BeansLive The average number of bean objects of this
specific type that are instantiated but not yet
destroyed. This applies to an individual bean,
regardless of its type.
StatelessBeansLive The average number of stateless session bean
objects that are instantiated but not yet destroyed
StatefulBeansLive The average number of stateful session bean
objects that are instantiated but not yet destroyed
EntityBeansLive The average number of entity bean objects that
are instantiated but not yet destroyed
BeanMethodRT The average method response time for all
methods defined in the remote interface to this
bean. Applies to all beans
BeanMethodActive The average number of methods being processed
concurrently. Applies to all beans
BeanMethodCalls The total number of method calls against this
bean
246
BeanObjectPool
The server holds a cache of bean objects. Each home has a cache and there is
therefore one BeanObjectPoolContainer per container. The top level
BeanObjectPool holds an aggregate of all the containers data.
OrbThreadPool
These are resources related to the ORB thread pool that is on the server.
BeanObjectPoolContainer The pool of a specific bean type
BeanObject The pool specific to a home
NumGet The number of calls retrieving an object from the
pool
NumGetFound The number of calls to the pool that resulted in
finding an available bean
NumPuts The number of beans that were released to the
pool
NumPutsDiscarded The number of times releasing a bean to the pool
resulted in the bean being discarded because the
pool was full
NumDrains The number of times the daemon found the pool
was idle and attempted to clean it
DrainSize The average number of beans discarded by the
daemon during a clean
BeanPoolSize The average number of beans in the pool
ActiveThreads The average number of active threads in the pool
TotalThreads The average number of threads in the pool
PercentTimeMaxed The average percentage of the time that the number
of threads in the pool reached or exceeded the
desired maximum number
247
DBConnectionMgr
These are resources related to the database connection manager. The
manager consists of a series of data sources, as well as a top-level aggregate
of each of the performance metrics.
ThreadCreates The number of threads created
ThreadDestroys The number of threads destroyed
ConfiguredMaxSize The configured maximum number of pooled
threads
DataSource Resources related to a specific data source specified
by the "name" attribute
ConnectionCreates The number of connections created
ConnectionDestroys The number of connections released
ConnectionPoolSize The average size of the pool, i.e., number of
connections
ConnectionAllocates The number of times a connection was allocated
ConnectionWaiters The average number of threads waiting for a
connection
ConnectionWaitTime The average time, in seconds, of a connection
grant
ConnectionTime The average time, in seconds, that a connection is
in use
ConnectionPercentUsed The average percentage of the pool that is in use
ConnectionPercentMaxed The percentage of the time that all connections are
in use
248
TransactionData
These are resources that pertain to transactions.
ServletEngine
These are resources that are related to servlets and JSPs.
NumTransactions The number of transactions processed
ActiveTransactions The average number of active transactions
TransactionRT The average duration of each transaction
BeanObjectCount The average number of bean object pools involved
in a transaction
RolledBack The number of transactions rolled back
Commited The number of transactions committed
LocalTransactions The number of transactions that were local
TransactionMethodCount The average number of methods invoked as part of
each transaction
Timeouts The number of transactions that timed out due to
inactivity timeouts
TransactionSuspended The average number of times that a transaction was
suspended
ServletsLoaded The number of servlets currently loaded
ServletRequests The number of requests serviced
CurrentRequests The number of requests currently being serviced
ServletRT The average response time for each request
ServletsActive The average number of servlets actively processing
requests
249
Sessions
These are general metrics regarding the HTTP session pool.
ServletIdle The amount of time that the server has been idle
(i.e., time since last request)
ServletErrors The number of requests that resulted in an error or
an exception
ServletBeanCalls The number of bean method invocations that were
made by the servlet
ServletBeanCreates The number of bean references that were made by
the servlet
ServletDBCalls The number of database calls made by the servlet
ServletDBConAlloc The number of database connections allocated by
the servlet
SessionLoads The number of times the servlet session data was
read from the database
SessionStores The number of times the servlet session data was
stored in the database
SessionSize The average size, in bytes, of a session data
LoadedSince The time that has passed since the server was loaded
(UNC time)
SessionsCreated The number of sessions created on the server
SessionsActive The number of currently active sessions
SessionsInvalidated The number of invalidated sessions. May not be
valid when using sessions in the database mode
SessionLifetime Contains statistical data of sessions that have been
invalidated. Does not include sessions that are still
alive
250
251
31
WebSphere Application Server
Monitoring
The WebSphere Application Server monitor allows you to monitor the
availability and server statistics of a IBM WebSphere Application Server
3.5.x, 4.x, and 5.x. The error and warning thresholds for the monitor can be
set on as many as ten WebSphere Application Server performance statistics.
The WebSphere Application Server monitor monitors the server
performance statistics from IBM WebSphere servers using the performance
monitoring interfaces provided with WebSphere. You can monitor multiple
parameters or counters with a single monitor instance. This allows you to
watch server loading for performance, availability, and capacity planning.
Create a separate WebSphere Application Server monitor instance for each
WebSphere application server in your environment.
To obtain data for this graph, you need to configure the WebSphere
Application Server monitor and select the measurements you want to
display before running the load test.
Setting up the Monitoring Environment on page 252
Configuring the WebSphere Application Server Monitor on page 254
Chapter 31 WebSphere Application Server Monitoring
252
Setting up the Monitoring Environment
Before you can use the WebSphere Application Server monitor, you need to
configure the server environment.
For WebSphere 3.5.x and 4.x
You must first install the IBM WebSphere Administrator's Console on the
SiteScope server if you are monitoring WebSphere versions 3.5.x or 4.x.
When installing the Administrator's Console:
Select the Custom installation option.
In the Choose Application Server Components dialog box, select
Administrator's Console and IBM JDK 1.2.2.
You will need to specify the machine you want to monitor during the
installation.
You must enable the WebSphere servers to be monitored.
For WebSphere 3.5.x, enable EPM Counters on the WebSphere server.
For WebSphere 4.x and 5.x, enable PMI Counters or enable the
Performance Monitoring Service on the WebSphere server.
You enable the counters for the application you want to monitor through
the WebSphere Administrator's Console.
For WebSphere 4.x:
When you have selected resources to monitor, select the Performance
option.
In the dialog box that opens, expand the Performance Modules tree. In
order to manage different levels of performance data, select the
performance modules and select a performance level, and click Set.
Alternatively, on WebSphere 3.5.x, you can set the EPM Specification to:
epm=high:epm.beanMethodData=none
253
through the WebSphere Administrator's Console.
If security has been enabled on the WebSphere server, the server security
ring must be copied to the admin client.
For WebSphere 5.x
Note: To monitor WebSphere version 5.x, the necessary WebSphere libraries
must be available on the SiteScope server. This means that a WebSphere 5.x
client must be installed on the SiteScope server.
To configure the server environment:
1 Select Servers > Application Servers, and select the server to be monitored
from the Application Server list.
2 In the Configuration tabs Additional Properties list, click Performance
Monitoring Service.
3 Select the Start Up check box.
4 In the Initial specification level section select Standard or Custom.
5 Click Apply.
To install the correct client software on a SiteScope server:
1 When installing WebSphere 5.x, select the following options from the
Custom Options menu:
Administration (or admin console)
Performance Analysis
Note: Certain trial versions of IBM WebSphere do not include the
Performance Analysis option required by the Sitescope WebSphere
Application Server monitor. The SiteScope monitor will only work when a
complete WebSphere production installation is available.
254
2 Copy all of the files from the <WebSphere 5.x Application Server
installation>\lib folder to the <client installation>\lib folder (see step 1
above).
3 The WebSphere 5.x server and client settings have to match. This means
that the SiteScope WebSphere Application Server Monitor will not be able to
monitor a WebSphere 5.1 application server if the client libraries are from a
WebSphere 5.0 application server, and vice versa.
Client libraries should be installed in separate folders with clearly distinct
directory names, such as WebSphere50 and WebSphere51, to avoid
confusion and SiteScope setup errors.
4 The sas.props file should be replaced with soap.props for WebSphere 5.x
installations.
Note: The WebSphere 5.x SiteScope monitor uses the WebSphere JMX
interface, so the port number used to communicate with the application
server is the SOAP port number. By default, the SOAP port number is 8880.
5 If security has been enabled on the WebSphere server, the server security
ring must be copied to the admin client.
General information
If security has been enabled on the WebSphere server, you must copy the
security keyring from the WebSphere server to SiteScope. A keyring is a
certification used by the server to identify the client.
Configuring the WebSphere Application Server Monitor
After you have configured the server environment, you need to configure
the WebSphere Application Server monitor and select the measurements
you want to display.
To configure the WebSphere Application Server monitor:
255
Profile.
Click Monitors.
2 Under Web Application Server Resource Graphs, select WebSphere
Application Server (SiteScope) and click Next. The Choose Server page
opens.
256
server whose resources you want to monitor.
port (default: 8888), and specify whether you are using a Secure HTTP
connection or a SiteScope user account. If you are using a SiteScope account,
fill in the account name or number, and the user name and password
defined to log in to the SiteScope account.
4 Click Next. The Configuring WebSphere Application Server monitor page
opens.
Type values for the monitor properties:
Client Properties File. Type the name of the custom client properties file.
For WebSphere 5.x+, you should select an appropriate soap.client.props
file. By default, the /properties/soap.client.props file is used.
Version. Select the version of the WebSphere server.
257
WebSphere Directory. Type the path of the WebSphere directory. This
directory should contain at least an Admin Console installation.
Port Number. Type the port number of the WebSphere server. This
should be the SOAP port for WebSphere 5.x+. The default port number is
8880.
Classpath. Type any extra classpath elements needed for the monitor
program.
Password. Type the password that SiteScope should use to log on to
WebSphere server.
Security Realm. Type the security realm for the WebSphere server (3.5x
only).
User Name. Type the user name that SiteScope should use to log on to
WebSphere server.
Server. Type the name of the server where the WebSphere application is
running. Do not type back-slashes (\\) that indicate a UNC path as part
of the name of the server.
Update every. The number entered here indicates the amount of time, in
seconds, between one monitor check and the next. By default the
monitor updates every 30 seconds.
258
6 Select the measurements you want to monitor. There is a large number of
counters available for the WebSphere Application Server Monitor. The list of
available counters will vary depending on which version of WebSphere you
are running.
For a list of counters that may be available, see WebSphere Performance
profile.
259
32
WebSphere (EPM) Monitoring
The WebSphere (EPM) monitor displays statistics about the resource usage
on the WebSphere 3.5.x server during the load test run.
Configuring the WebSphere (EPM) Monitor on page 267
To monitor the IBM WebSphere application server (3.5.x), you must first
install the IBM WebSphere Administrators Console on the Performance
Center Host machine and enable the EPM counters. You can then configure
the WebSphere (EPM) monitor before running the load test.
Installing the IBM WebSphere Administrators Console on page 260
Copying the Security Keyring on page 265
Enabling EPM Counters on the WebSphere 3.5.x Server on page 266
Chapter 32 WebSphere (EPM) Monitoring
260
Installing the IBM WebSphere Administrators Console
You must first install the IBM WebSphere Administrators Console on the
Performance Center Host machine.
To install the IBM WebSphere Administrators Console:
1 Start the WebSphere installation program from the WebSphere 3.5 Windows
NT distribution CD-ROM. The WebSphere Application Server dialog box
opens.
2 Disregard the instruction to shut down all Web servers that you plan to run
with WebSphere. This is not relevant to the Administrators Console
installation. Follow the remaining instructions.
261
3 Click Next to proceed. The Installation Options dialog box opens.
262
4 Select Custom Installation, and click Next. The Choose Application Server
Components dialog box opens.
5 Select Administrators Console and IBM JDK 1.2.2. Clear all the other
options.
6 Click Next. The Get Host Name dialog box opens.
263
7 Type the name of the machine that you want to monitor.
8 Click Next. The Product Directory dialog box opens.
9 Specify the folder in which to install the Administrators Console. To select a
different location, click Browse, select a folder other than the default folder,
and click OK.
10 Click Next. The Select Program Folder dialog box opens.
264
11 Specify a program folder, or accept the default folder,
IBM WebSphere\Application Server V3.5.
12 Click Next. The installation process begins. To pause or quit the installation,
click Cancel.
When the installation is complete, the Setup Complete dialog box opens.
13 In the Setup Complete dialog box, select the check box to view the readme
file before starting the program. You can view the readme file at any time by
selecting Start > Programs > Application Server V3.5 > IBM WebSphere >
README.
265
14 Click Finish to complete the installation program. The Restarting Windows
dialog box opens.
15 Select either to restart your computer and complete the installation now
(recommended) or to wait and complete the installation later.
16 Click OK to complete the installation of the Administrators Console.
Copying the Security Keyring
If you enabled security on the WebSphere server, you must copy the security
keyring from the server to the admin client. (One way to tell whether
security is enabled is to see whether the Administrators Console can
connect to the admin server.) A keyring is a certification used by the server
to identify the client.
You need to copy the jar file containing the keyring from the server lib
folder to the client lib folder. You also need to add the jar file containing the
keyring to the monitoring client command line.
Note: The keyring used in this file (353Keyring.jar) is the IBM dummy
keyring that must be installed on servers using versions 3.52 and below. If
your server is using the IBM dummy keyring and is version 3.52 or below,
you do not need to change the line. If you are using the dummy keyring and
are running version 3.53 or later, you do not need to do anything.
266
To copy the keyring:
1 Copy the keyring jar file from the server to the admin client lib folder (by
default, C:\Websphere\Appserver\lib):
The jar file containing the keyring, xxxKeyring.jar, is located by default in
the following location:
NT Server. C:\Websphere\Appserver\lib
UNIX Server. OPT/websphere/Appserver/lib
2 Open the <Performance Center root
folder>\dat\monitors\WebSphere35Mon.ini file in a text editor.
3 Locate the following line:
Note: If you did not use the default location for the WebSphere installation,
the line will be different.
4 Change 353Keyring.jar to the keyring you are using.
Enabling EPM Counters on the WebSphere 3.5.x Server
To enable the EPM counters, which are by default set to "none," right-click
the application you are monitoring in the WebSphere Administrators
Console browser, and select Performance. Expand the Performance Modules
tree in the dialog box that opens. To manage different levels of performance
data, right-click the performance modules and select a performance level.
Click the Set button.
Alternatively, check that the application server is started, select the
Advanced tab in the WebSphere Administrators Console browser, and in
the EPM Specification box, type: epm=high:epm.beanMethodData=none
JVM_CLASSES4=C:\WebSphere\AppServer\lib\353Keyring.ja
267
Configuring the WebSphere (EPM) Monitor
Once you have installed the WebSphere Administrators Console and
enabled the EPM counters, you select measurements to monitor the
WebSphere EPM server using the WebSphere EPM Choose Measurements
page.
To activate the WebSphere EPM monitor:
Profile.
Click Monitors.
2 Under Web Application Server Resource Graphs, select WebSphere EPM and
monitor.
268
For a list of the available performance counters, see WebSphere
profile.
Part IX
Database Resource Monitoring
270
271
33
Introducing Database Resource
Monitoring
To isolate server performance bottlenecks, you can monitor the database
server resource usage of an SQL, DB2, Sybase, and Oracle database server
during a load test run.
To obtain performance data, you need to activate the online monitor for the
server and specify which resources you want to measure before running the
load test.
For more information, see:
Chapter 34, SQL Server Monitoring
Chapter 35, DB2 Database Server Monitoring
Chapter 36, Oracle Database Server Monitoring
Chapter 37, Sybase Database Server Monitoring
Chapter 33 Introducing Database Resource Monitoring
272
273
34
SQL Server Monitoring
The SQL Server monitor measures the standard Windows resources on the
SQL server machine.
Configuring the SQL Server Monitor on page 273
SQL Server Performance Counters on page 276
Configuring the SQL Server Monitor
To obtain SQL Server performance data, you need to activate the online
monitor for the server and specify which resources you want to measure
before running the load test. You select measurements to monitor using the
SQL Server Choose Measurements page.
We recommend using the SiteScope monitor engine. In this case, SiteScope
must be installed on a server. Although SiteScope is installed on the Utility
Server, we recommend installing it on the same machine as the Controller,
or on a dedicated server in order to reduce load on the Utility Server.
Note: To monitor an SQL server through a firewall, use TCP, port 139.
To configure the SQL server monitor:
Chapter 34 SQL Server Monitoring
274
Profile.
Click Monitors.
2 Under Database Server Resource Graphs, select SQL Server or SQL Server
(SiteScope) and click Next. The Choose Server page opens.
3 Type the name or IP address of the SQL server whose resources you want to
monitor.
For SQL Server, type the user login name and password on the SQL server.
For SQL Server (SiteScope), type the SiteScope machine name and port
(default: 8888), and specify whether you are using a Secure HTTP
Click Next.
275
4 The first time you add measurements using the SQL Server (SiteScope)
monitor you need to configure the NT remote machine properties. Enter the
remote machines configuration information, as described in Configuring
the NT Remote Machine on page 27.
For a list of the available performance counters, see SQL Server
profile.
Note: Certain measurements or counters are especially useful for
determining server performance and isolating the cause of a bottleneck
during an initial stress test on the SQL Server. For more information about
these counters, see the section that describes useful counters for stress
testing in the HP LoadRunner Analysis User Guide.
276
SQL Server Performance Counters
The following table describes some of the counters that can be measured on
version 6.5 of the SQL Server:
% Total Processor
Time (NT)
The average percentage of time that all the processors on
the system are busy executing non-idle threads. On a
multi-processor system, if all processors are always busy,
this is 100%, if all processors are 50% busy this is 50%
and if 1/4th of the processors are 100% busy this is 25%.
It can be viewed as the fraction of the time spent doing
useful work. Each processor is assigned an Idle thread in
the Idle process which consumes those unproductive
processor cycles not used by any other threads.
% Processor Time
(Win 2000)
The percentage of time that the processor is executing a
non-idle thread. This counter was designed as a primary
indicator of processor activity. It is calculated by
measuring the time that the processor spends executing
the thread of the idle process in each sample interval,
and subtracting that value from 100%. (Each processor
has an idle thread which consumes cycles when no other
threads are ready to run). It can be viewed as the
percentage of the sample interval spent doing useful
work. This counter displays the average percentage of
busy time observed during the sample interval. It is
calculated by monitoring the time the service was
inactive, then subtracting that value from 100%.
Cache Hit Ratio The percentage of time that a requested data page was
found in the data cache (instead of being read from disk)
I/O - Batch
Writes/sec
The number of pages written to disk per second, using
Batch I/O. The checkpoint thread is the primary user of
Batch I/O.
I/O - Lazy Writes/sec The number of pages flushed to disk per second by the
Lazy Writer
I/O - Outstanding
Reads
The number of physical reads pending
277
I/O - Outstanding
Writes
The number of physical writes pending
I/O - Page Reads/sec The number of physical page reads per second
I/O -
Transactions/sec
The number of Transact-SQL command batches executed
per second
User Connections The number of open user connections
278
279
35
DB2 Database Server Monitoring
The DB2 Database Server monitor measures the resource usage on a DB2
database during a load test run.
Note: If there is no application working with a database, you can only
monitor the database manager instance.
Configuring the DB2 Monitor on page 281
DB2 Performance Counters on page 282
Before you can monitor a DB2 database server, you must set up the DB2
monitor environment.
To set up the DB2 monitor environment:
1 Install all the client files and libraries on the Performance Center Host
machine.
2 Install all the client files and libraries on the Utility Server.
3 Select Start > Programs > DB2 for Windows NT > Control Center. Type your
DB2 server user name and password (with administrative privileges).
4 In the console that opens, right-click Systems, and select Add.
Chapter 35 DB2 Database Server Monitoring
280
5 Enter the following settings in the dialog box:
System Name. The server name.
Remote Instance. DB2.
Host Name. The host server name.
Service Name. The DB2 server port. The default value is 50000.
6 Click Retrieve and OK.
Note: If you receive an error message after clicking Retrieve, repeat steps 4
and 5, and click OK.
7 Expand the <server name> node in the console tree.
8 Right-click Instance, and select Add.
9 Enter the following settings in the dialog box:
Remote Instance. DB2
Instance Name. The database instance to be called from the Controller.
Host Name. The host server name.
Service Name. The DB2 server port. The default value is 50000.
10 Click OK and close the Control Center.
Note: You can only work with a single Database Manager instance during
each monitoring session.
281
Configuring the DB2 Monitor
To obtain performance data, you need to activate the DB2 monitor for the
server and specify which resources you want to measure before running the
load test. You select measurements to monitor using the DB2 Choose
Measurements page.
To configure the DB2 monitor:
Profile.
Click Monitors.
2 Under Database Server Resource Graphs, select DB2 and click Next. The
monitor.
282
For a list of the available performance counters, see DB2 Performance
profile.
DB2 Performance Counters
The following sections describe the default counters that can be monitored:
DatabaseManager Counters on page 283
Database Counters on page 284
Application Counters on page 291
283
DatabaseManager Counters
rem_cons_in The current number of connections initiated
from remote clients to the instance of the
database manager that is being monitored.
rem_cons_in_exec The number of remote applications that are
currently connected to a database and are
currently processing a unit of work within the
database manager instance being monitored.
local_cons The number of local applications that are
currently connected to a database within the
database manager instance being monitored.
local_cons_in_exec The number of local applications that are
currently connected to a database within the
database manager instance being monitored and
are currently processing a unit of work.
con_local_dbases The number of local databases that have
applications connected.
agents_registered The number of agents registered in the database
manager instance that is being monitored
(coordinator agents and subagents).
agents_waiting_on_token The number of agents waiting for a token so they
can execute a transaction in the database
manager.
idle_agents The number of agents in the agent pool that are
currently unassigned to an application and are
therefore "idle".
agents_from_pool The number of agents assigned from the agent
pool
agents_created_empty_pool The number of agents created because the agent
pool was empty.
284
Database Counters
agents_stolen The number of times that agents are stolen from
an application. Agents are stolen when an idle
agent associated with an application is reassigned
to work on a different application.
comm_private_mem The amount of private memory that the instance
of the database manager has currently committed
at the time of the snapshot.
inactive_gw_agents The number of DRDA agents in the DRDA
connections pool that are primed with a
connection to a DRDA database, but are inactive.
num_gw_conn_switches The number of times that an agent from the
agents pool was primed with a connection and
was stolen for use with a different DRDA
database.
sort_heap_allocated The total number of allocated pages of sort heap
space for all sorts at the level chosen and at the
time the snapshot was taken.
post_threshold_sorts The number of sorts that have requested heaps
after the sort heap threshold has been reached.
piped_sorts_requested The number of piped sorts that have been
requested.
piped_sorts_accepted The number of piped sorts that have been
accepted.
Measurement
Description
appls_cur_cons Indicates the number of applications that are
currently connected to the database.
appls_in_db2 Indicates the number of applications that are
currently connected to the database, and for
which the database manager is currently
processing a request.
285
total_sec_cons The number of connections made by a sub-agent
to the database at the node.
num_assoc_agents At the application level, this is the number of sub-
agents associated with an application. At the
database level, it is the number of sub-agents for
all applications.
sort_heap_allocated The total number of allocated pages of sort heap
space for all sorts at the level chosen and at the
time the snapshot was taken.
total_sorts The total number of sorts that have been
executed.
total_sort_time The total elapsed time (in milliseconds) for all
sorts that have been executed.
sort_overflows The total number of sorts that ran out of sort
heap and may have required disk space for
temporary storage.
active_sorts The number of sorts in the database that
currently have a sort heap allocated.
total_hash_joins The total number of hash joins executed.
total_hash_loops The total number of times that a single partition
of a hash join was larger than the available sort
heap space.
hash_join_overflows The number of times that hash join data exceeded
the available sort heap space
hash_join_small_overflows The number of times that hash join data exceeded
the available sort heap space by less than 10%.
pool_data_l_reads Indicates the number of logical read requests for
data pages that have gone through the buffer
pool.
pool_data_p_reads The number of read requests that required I/O to
get data pages into the buffer pool.
Measurement
Description
286
pool_data_writes Indicates the number of times a buffer pool data
page was physically written to disk.
pool_index_l_reads Indicates the number of logical read requests for
index pages that have gone through the buffer
pool.
pool_index_p_reads Indicates the number of physical read requests to
get index pages into the buffer pool.
pool_index_writes Indicates the number of times a buffer pool index
pool_read_time Provides the total amount of elapsed time spent
processing read requests that caused data or index
pages to be physically read from disk to buffer
pool.
pool_write_time Provides the total amount of time spent
physically writing data or index pages from the
buffer pool to disk.
files_closed The total number of database files closed.
pool_async_data_reads The number of pages read asynchronously into
the buffer pool.
pool_async_data_writes The number of times a buffer pool data page was
physically written to disk by either an
asynchronous page cleaner, or a pre-fetcher. A
pre-fetcher may have written dirty pages to disk
to make space for the pages being pre-fetched.
pool_async_index_writes The number of times a buffer pool index page was
physically written to disk by either an
asynchronous page cleaner, or a pre-fetcher. A
pre-fetcher may have written dirty pages to disk
to make space for the pages being pre-fetched.
pool_async_index_reads The number of index pages read asynchronously
into the buffer pool by a pre-fetcher.
pool_async_read_time The total elapsed time spent reading by database
manager pre-fetchers.
Measurement
Description
287
pool_async_write_time The total elapsed time spent writing data or index
pages from the buffer pool to disk by database
manager page cleaners.
pool_async_data_read_reqs The number of asynchronous read requests.
pool_lsn_gap_clns The number of times a page cleaner was invoked
because the logging space used had reached a pre-
defined criterion for the database.
pool_drty_pg_steal_clns The number of times a page cleaner was invoked
because a synchronous write was needed during
the victim buffer replacement for the database.
pool_drty_pg_thrsh_clns The number of times a page cleaner was invoked
because a buffer pool had reached the dirty page
threshold criterion for the database.
prefetch_wait_time The time an application spent waiting for an I/O
server (pre-fetcher) to finish loading pages into
the buffer pool.
pool_data_to_estore The number of buffer pool data pages copied to
extended storage.
pool_index_to_estore The number of buffer pool index pages copied to
extended storage.
pool_data_from_estore The number of buffer pool data pages copied
from extended storage.
pool_index_from_estore The number of buffer pool index pages copied
direct_reads The number of read operations that do not use
the buffer pool.
direct_writes The number of write operations that do not use
the buffer pool.
direct_read_reqs The number of requests to perform a direct read
of one or more sectors of data.
direct_write_reqs The number of requests to perform a direct write
Measurement
Description
288
direct_read_time The elapsed time (in milliseconds) required to
perform the direct reads.
direct_write_time The elapsed time (in milliseconds) required to
perform the direct writes.
cat_cache_lookups The number of times that the catalog cache was
referenced to obtain table descriptor information.
cat_cache_inserts The number of times that the system tried to
insert table descriptor information into the
catalog cache.
cat_cache_overflows The number of times that an insert into the
catalog cache failed due the catalog cache being
full.
cat_cache_heap_full The number of times that an insert into the
catalog cache failed due to a heap-full condition
in the database heap.
pkg_cache_lookups The number of times that an application looked
for a section or package in the package cache. At a
database level, it indicates the overall number of
references since the database was started, or
monitor data was reset.
pkg_cache_inserts The total number of times that a requested
section was not available for use and had to be
loaded into the package cache. This count
includes any implicit prepares performed by the
system.
pkg_cache_num_overflows The number of times that the package cache
overflowed the bounds of its allocated memory.
appl_section_lookups Lookups of SQL sections by an application from
its SQL work area.
appl_section_inserts Inserts of SQL sections by an application from its
SQL work area.
sec_logs_allocated The total number of secondary log files that are
currently being used for the database.
Measurement
Description
289
log_reads The number of log pages read from disk by the
logger.
log_writes The number of log pages written to disk by the
logger.
total_log_used The total amount of active log space currently
used (in bytes) in the database.
locks_held The number of locks currently held.
lock_list_in_use The total amount of lock list memory (in bytes)
that is in use.
deadlocks The total number of deadlocks that have
occurred.
lock_escals The number of times that locks have been
escalated from several row locks to a table lock.
x_lock_escals The number of times that locks have been
escalated from several row locks to one exclusive
table lock, or the number of times an exclusive
lock on a row caused the table lock to become an
exclusive lock.
lock_timeouts The number of times that a request to lock an
object timed-out instead of being granted.
lock_waits The total number of times that applications or
connections waited for locks.
lock_wait_time The total elapsed time waited for a lock.
locks_waiting Indicates the number of agents waiting on a lock.
rows_deleted The number of row deletions attempted.
rows_inserted The number of row insertions attempted.
rows_updated The number of row updates attempted.
rows_selected The number of rows that have been selected and
returned to the application.
Measurement
Description
290
int_rows_deleted The number of rows deleted from the database as
a result of internal activity.
int_rows_updated The number of rows updated from the database as
int_rows_inserted The number of rows inserted into the database as
a result of internal activity caused by triggers.
static_sql_stmts The number of static SQL statements that were
attempted.
dynamic_sql_stmts The number of dynamic SQL statements that
were attempted.
failed_sql_stmts The number of SQL statements that were
attempted, but failed.
commit_sql_stmts The total number of SQL COMMIT statements
that have been attempted.
rollback_sql_stmts The total number of SQL ROLLBACK statements
select_sql_stmts The number of SQL SELECT statements that were
executed.
uid_sql_stmts The number of SQL UPDATE, INSERT, and
DELETE statements that were executed.
ddl_sql_stmts This element indicates the number of SQL Data
Definition Language (DDL) statements that were
executed.
int_auto_rebinds The number of automatic rebinds (or recompiles)
int_commits The total number of commits initiated internally
by the database manager.
int_rollbacks The total number of rollbacks initiated internally
Measurement
Description
291
Application Counters
int_deadlock_rollbacks The total number of forced rollbacks initiated by
the database manager due to a deadlock. A
rollback is performed on the current unit of work
in an application selected by the database
manager to resolve the deadlock.
binds_precompiles The number of binds and pre-compiles
attempted.
Measurement
Description
agents_stolen The number of times that agents are stolen from
an application. Agents are stolen when an idle
agent associated with an application is reassigned
to work on a different application.
num_assoc_agents At the application level, this is the number of sub-
agents associated with an application. At the
database level, it is the number of sub-agents for
all applications.
total_sorts The total number of sorts that have been
executed.
total_sort_time The total elapsed time (in milliseconds) for all
sorts that have been executed.
sort_overflows The total number of sorts that ran out of sort
heap and may have required disk space for
temporary storage.
total_hash_joins The total number of hash joins executed.
total_hash_loops The total number of times that a single partition
of a hash join was larger than the available sort
heap space.
hash_join_overflows The number of times that hash join data exceeded
the available sort heap space
Measurement
Description
292
hash_join_small_overflows The number of times that hash join data exceeded
the available sort heap space by less than 10%.
pool_data_l_reads Indicates the number of logical read requests for
data pages that have gone through the buffer
pool.
pool_data_p_reads The number of read requests that required I/O to
get data pages into the buffer pool.
pool_data_writes Indicates the number of times a buffer pool data
pool_index_l_reads Indicates the number of logical read requests for
index pages that have gone through the buffer
pool.
pool_index_p_reads Indicates the number of physical read requests to
get index pages into the buffer pool.
pool_index_writes Indicates the number of times a buffer pool index
pool_read_time Provides the total amount of elapsed time spent
processing read requests that caused data or index
pages to be physically read from disk to buffer
pool.
prefetch_wait_time The time an application spent waiting for an I/O
server (pre-fetcher) to finish loading pages into
the buffer pool.
pool_data_to_estore The number of buffer pool data pages copied to
extended storage.
pool_index_to_estore The number of buffer pool index pages copied to
extended storage.
pool_data_from_estore The number of buffer pool data pages copied
pool_index_from_estore The number of buffer pool index pages copied
Measurement
Description
293
direct_reads The number of read operations that do not use
the buffer pool.
direct_writes The number of write operations that do not use
the buffer pool.
direct_read_reqs The number of requests to perform a direct read
direct_write_reqs The number of requests to perform a direct write
direct_read_time The elapsed time (in milliseconds) required to
perform the direct reads.
direct_write_time The elapsed time (in milliseconds) required to
perform the direct writes.
cat_cache_lookups The number of times that the catalog cache was
referenced to obtain table descriptor information.
cat_cache_inserts The number of times that the system tried to
insert table descriptor information into the
catalog cache.
cat_cache_overflows The number of times that an insert into the
catalog cache failed due the catalog cache being
full.
cat_cache_heap_full The number of times that an insert into the
catalog cache failed due to a heap-full condition
in the database heap.
pkg_cache_lookups The number of times that an application looked
for a section or package in the package cache. At a
database level, it indicates the overall number of
references since the database was started, or
monitor data was reset.
pkg_cache_inserts The total number of times that a requested
section was not available for use and had to be
loaded into the package cache. This count
includes any implicit prepares performed by the
system.
Measurement
Description
294
appl_section_lookups Lookups of SQL sections by an application from
its SQL work area.
appl_section_inserts Inserts of SQL sections by an application from its
SQL work area.
uow_log_space_used The amount of log space (in bytes) used in the
current unit of work of the monitored
application.
locks_held The number of locks currently held.
deadlocks The total number of deadlocks that have
occurred.
lock_escals The number of times that locks have been
escalated from several row locks to a table lock.
x_lock_escals The number of times that locks have been
escalated from several row locks to one exclusive
table lock, or the number of times an exclusive
lock on a row caused the table lock to become an
exclusive lock.
lock_timeouts The number of times that a request to lock an
object timed-out instead of being granted.
lock_waits The total number of times that applications or
connections waited for locks.
lock_wait_time The total elapsed time waited for a lock.
locks_waiting Indicates the number of agents waiting on a lock.
uow_lock_wait_time The total amount of elapsed time this unit of
work has spent waiting for locks.
rows_deleted The number of row deletions attempted.
rows_inserted The number of row insertions attempted.
rows_updated The number of row updates attempted.
rows_selected The number of rows that have been selected and
returned to the application.
Measurement
Description
295
rows_written The number of rows changed (inserted, deleted or
updated) in the table.
rows_read The number of rows read from the table.
int_rows_deleted The number of rows deleted from the database as
int_rows_updated The number of rows updated from the database as
int_rows_inserted The number of rows inserted into the database as
a result of internal activity caused by triggers.
open_rem_curs The number of remote cursors currently open for
this application, including those cursors counted
by open_rem_curs_blk.
open_rem_curs_blk The number of remote blocking cursors currently
open for this application.
rej_curs_blk The number of times that a request for an I/O
block at server was rejected and the request was
converted to non-blocked I/O.
acc_curs_blk The number of times that a request for an I/O
block was accepted.
open_loc_curs The number of local cursors currently open for
this application, including those cursors counted
by open_loc_curs_blk.
open_loc_curs_blk The number of local blocking cursors currently
open for this application.
static_sql_stmts The number of static SQL statements that were
attempted.
dynamic_sql_stmts The number of dynamic SQL statements that
were attempted.
failed_sql_stmts The number of SQL statements that were
attempted, but failed.
Measurement
Description
296
commit_sql_stmts The total number of SQL COMMIT statements
rollback_sql_stmts The total number of SQL ROLLBACK statements
select_sql_stmts The number of SQL SELECT statements that were
executed.
uid_sql_stmts The number of SQL UPDATE, INSERT, and
DELETE statements that were executed.
ddl_sql_stmts This element indicates the number of SQL Data
Definition Language (DDL) statements that were
executed.
int_auto_rebinds The number of automatic rebinds (or recompiles)
int_commits The total number of commits initiated internally
int_rollbacks The total number of rollbacks initiated internally
int_deadlock_rollbacks The total number of forced rollbacks initiated by
the database manager due to a deadlock. A
rollback is performed on the current unit of work
in an application selected by the database
manager to resolve the deadlock.
binds_precompiles The number of binds and pre-compiles
attempted.
Measurement
Description
297
36
Oracle Database Server Monitoring
The Oracle database server resource monitor measures statistics for Oracle
database servers. During a test run, you use this monitor to isolate Oracle
database server performance bottlenecks.
The Oracle monitor displays information from Oracle V$ tables: Session
statistics, V$SESSTAT, system statistics, V$SYSSTAT, and other table counters
defined by the user in the custom query.
The SiteScope Oracle JDBC monitor monitors the server performance
statistics from Oracle database servers. You can monitor multiple parameters
or counters with a single monitor instance. This allows you to watch server
loading for performance, availability, and capacity planning. You can create
a separate Oracle JDBC Monitor instance for each Oracle database server in
your environment.
Note: The port you use to monitor an Oracle server through a firewall
depends on the configuration of the Oracle server. Configuration
information for the connection between the client and server is located in
the Oracle client tnsnames.ora file.
Setting Up the Oracle Monitor Environment on page 298
Configuring Oracle Database Resource Monitoring on page 302
Oracle Performance Counters on page 306
Custom Queries on page 307
Chapter 36 Oracle Database Server Monitoring
298
Setting Up the Oracle Monitor Environment
The Oracle server measures information from the V$SESSTAT and
V$SYSSTAT Oracle V$ tables, and other table counters defined by the user in
the custom query. In order to monitor the Oracle server, you must set up the
monitoring environment as described below before you can configure the
monitor.
To set up the SiteScope Oracle JDBC monitor environment:
1 You must have a copy of the applicable Oracle JDBC database driver file (for
example, classes12.zip) on the SiteScope server.
Copy the downloaded driver file into the <SiteScope install
path>\SiteScope\java\lib\ext subdirectory. DO NOT unzip the file.
Stop and restart the SiteScope service after copying the driver file to the
SiteScope machine.
Note: More than one driver file is available for download. Some drivers
support more than one version of Oracle database (for example, the
classes12.zip Oracle JDBC thin driver) while others only support a particular
version. If you are monitoring a recent version of Oracle database, you
should download the latest version of the database driver.
2 You must supply the correct Database Connection URL, a database user
name and password when setting up the monitor. The syntax of the
Database Connection URL usually has the form of:
jdbc:oracle:thin:@<tcp address>:<tcp port>:<database sid>
299
For example to connect to the ORCL database on a machine using port 1521
you would use:
Note: The colon and @ symbols must be included as shown.
3 You must specify the Oracle Database Driver that was installed on the
SiteScope server when setting up the monitor. The Database Driver for the
Oracle thin JDBC driver is:
4 You should only have one Oracle client installed on the SiteScope machine.
If there is more that one client installed, SiteScope may report an error and
be unable to connect to the database.
5 You must have an Oracle user login that SiteScope will use to access the
Oracle server. In order to retrieve the Oracle database counters, the user that
SiteScope will use for the Oracle JDBC Monitor should be able to execute all
the SQL statements as found in the file
SiteScope\templates.applications\commands.oraclejdbc.
To set up the Performance Center Oracle monitor environment:
1 Check that the Oracle client libraries are installed on the Performance
Center Host and Utility server machines.
2 Verify that %OracleHome%\bin is included in the path environment
variable. If it is not, add it.
3 Configure the tnsnames.ora file on the Performance Center Host machine
so that the Oracle client can communicate with the Oracle servers you plan
to monitor.
You can configure connection parameters either manually, by editing the
tnsnames.ora file in a text editor, or using the Oracle service configuration
tool (for example, select Start > Programs > Oracle for Windows NT > Oracle
Net8 Easy Config).
jdbc:oracle:thin:@206.168.191.19:1521:ORCL
oracle.jdbc.driver.OracleDriver
300
You specify:
a new service name (TNS name) for the Oracle instance
TCP protocol
the host name (name of monitored server machine)
the port number (usually 1521)
the database SID (the default SID is ORCL)
For example:
Note: Only the 32-bit Oracle client should be installed on the Performance
Center Host machine running the Oracle monitor. If you have a 16-bit and a
32-bit Oracle client installation on the Performance Center Host machine,
the 16-bit installation should be uninstalled.
4 Obtain a user name and password for the service from your database
administrator, and check that the Performance Center Host has database
administrator privileges for the Oracle V$ tables (V$SESSTAT, V$SYSSTAT,
V$STATNAME, V$INSTANCE, V$SESSION).
5 Verify connection with the Oracle server by performing tns ping from the
Performance Center Host machine. Note that there may be a problem
connecting if the Oracle server is behind a DMZ/firewall that limits its
communication to application servers accessing it.
301
6 Check that the registries are updated for the version of Oracle that you are
using and that they have the following key:
7 Verify that the Oracle server you want to monitor is up and running.
Note: It is possible to monitor several Oracle database servers concurrently.
8 Run SQL*Plus from each Performance Center Host machine and the Utility
server. Attempt to log in to the Oracle servers with the desired user
name/password/server combination.
9 Type SELECT * FROM V$SYSSTAT to verify that you can view the
V$SYSSTAT table on the Oracle server. Use similar queries to verify that you
can view the V$SESSTAT, V$SESSION, V$INSTANCE, V$STATNAME, and
V$PROCESS tables on the server. Make sure that the Oracle bin directory is
in the search path.
10 To change the length of each monitoring sample (in seconds), you need to
edit the dat\monitors\vmon.cfg file in the Performance Center root folder.
The default rate is 10 seconds.
Note: The minimum sampling rate for the Oracle Monitor is 10 seconds. If
you set the sampling rate at less than 10 seconds, the Oracle Monitor will
continue to monitor at 10 second intervals.
If a problem occurs in setting up the Oracle environment, view the error
message issued by the Oracle server.
HKEY_LOCAL_MACHINE\SOFTWARE\ORACLE
302
Configuring Oracle Database Resource Monitoring
After you have set up the Oracle monitor environment, you can select the
objects you want to measure.
To configure the Oracle monitor:
Profile.
Click Monitors.
2 Under Database Server Resource Graphs, select Oracle or Oracle (SiteScope)
and click Next. The Choose Server page opens.
For Oracle, type the user login name and password on the Oracle server.
For Oracle (SiteScope), type the SiteScope machine name and port
(default: 8888), and specify if you are using a Secure HTTP connection or
a SiteScope user account. If you are using a SiteScope account, fill in the
account name or number, and the user name and password defined to
log in to the SiteScope account.
Click Next.
303
4 For the Oracle (SiteScope) monitor, the Configuring Oracle JDBC monitor
page opens.
Enter the following Oracle JDBC monitor information, and click Next:
Instance. Type the database SID. For example, ORCL.
Database Connection UR. Type the connection URL to the database you
want to monitor. For example,
jdbc:oracle:thin:@206.168.191.19:1521:ORCL
Database User Name. Type the user name that SiteScope should use to
connect to the database.
Database Password. Type the password for the user name that SiteScope
should use to connect to the database.
Database Driver. Type the driver used to connect to the database. For
example, oracle.jdbc.driver.OracleDriver.
304
Connection Timeout. Type an optional time out value, in seconds, that
SiteScope should to wait for a database connection to respond. If the
database connection can not be completed within the period specified,
SiteScope will report an error.
Query Timeout. Type an optional the time out value, in seconds, that
SiteScope should to wait for a response from the database query. If the
database does not respond within the period specified, SiteScope will
report an error.
Update every. Select how often the monitor should read the server
statistics. The default interval is to run or update the monitor once every
30 seconds.
Note:
The sum of the Connection Timeout value and Query Timeout value
should always be less than the Update every value for the monitor.
Some commonly used databases and database drivers do not support the
Query Timeout feature. In these cases the Query Timeout value should be
set to zero.
305
6 For each measurement that you want to monitor, select an object,
measurements, and instances. The instances are relevant only if multiple
instances of the highlighted instance are running.
profile.
Note: By default, the database returns the absolute value of a counter.
However, by changing the IsRate setting in the dat\monitors\vmon.cfg file
to 1, you can instruct the database to report a counters rate valuethe
change in the counter per unit time.
306
Oracle Performance Counters
The following measurements are most commonly used when monitoring
the Oracle server (from the V$SYSSTAT table):
CPU used by this session This is the amount of CPU time (in 10s of
milliseconds) used by a session between the time a
user call started and ended. Some user calls can be
completed within 10 milliseconds and, as a result,
the start and end user-call time can be the same. In
this case, 0 milliseconds are added to the statistic. A
similar problem can exist in the operating system
reporting, especially on systems that suffer from
many context switches.
Bytes received via
SQL*Net from client
The total number of bytes received from the client
over Net8
Logons current The total number of current logons
Opens of replaced files The total number of files that needed to be
reopened because they were no longer in the
process file cache
User calls Oracle allocates resources (Call State Objects) to
keep track of relevant user call data structures every
time you log in, parse, or execute. When
determining activity, the ratio of user calls to RPI
calls gives you an indication of how much internal
work gets generated as a result of the type of
requests the user is sending to Oracle.
SQL*Net roundtrips
to/from client
The total number of Net8 messages sent to, and
received from, the client
Bytes sent via SQL*Net to
client
The total number of bytes sent to the client from
the foreground process(es)
Opened cursors current The total number of current open cursors
307
Custom Queries
Using the custom query feature, you can define your own query to the
Oracle database and view the result of this querya single numerical
valuein the Oracle online monitor graph. By defining your own query,
you can monitor the V$SYSSTAT and V$SESSTAT table counters that are
currently provided by the Oracle monitor, as well as other tables that
contain useful performance information.
To create a custom query:
1 In the third line of the vmon.cfg file, CustomCounters=, indicate the
number of custom counters you want to create.
2 Create a new section in the vmon.cfg file for the new counter. Each section
has the following format:
DB block changes Closely related to consistent changes, this statistic
counts the total number of changes that were made
to all blocks in the SGA that were part of an update
or delete operation. These are changes that are
generating redo log entries and hence will be
permanent changes to the database if the
transaction is committed. This statistic is a rough
indication of total database work and indicates
(possibly on a per-transaction level) the rate at
which buffers are being dirtied.
Total file opens The total number of file opens being performed by
the instance. Each process needs a number of files
(control file, log file, database file) to work against
the database.
[Custom2]
Name=Number of sessions
Description=This counter returns the number of sessions active.
Query=SELECT COUNT(*) FROM V$SESSION
IsRate=1
308
3 In the [Custom#] line, assign the next number in the sequence of counters
to the new custom counter. Note that the custom counters must be in
consecutive order, beginning with the number 0.
4 In the Name line, type the name of the new counter.
5 In the Description line, type the description of the counter that you want
the help message to contain.
6 In the Query line, type the text of the SQL query (on one line of the
vmon.cfg file) that returns exactly one row from the database. This row
must contain one column, a numerical value.
Note: Custom queries should not exceed 512 characters.
7 In the IsRate line, type 0 if you want the database to report the counter as an
absolute number. If you want the database to report the change in the
counter per unit time, type 1.
Note: Custom queries cannot return negative values.
309
37
Sybase Database Server Monitoring
The Sybase monitor enables monitoring of Sybase Adaptive Server
Enterprise (Sybase ASE) servers (version 11 or later) on Windows and UNIX.
The monitor connects to the Sybase ASE server (through the Adaptive Server
Enterprise Monitor Server) and retrieves metrics from the server using
standard, Sybase-provided libraries.
To monitor the Sybase database server machine, you must first set up the
Sybase monitoring environment. You then enable the Oracle online
monitor and select the default measurements you want to display, before
running the load test.
Note: When connecting to the monitored server, you connect to the
Adaptive Server Enterprise Monitor Server, not the Sybase ASE server. The
Adaptive Server Enterprise Monitor Server is an application that runs on the
same machine as Sybase ASE server and retrieves performance information
from it. The Adaptive Server Enterprise Monitor Server usually has the same
name as the Sybase server, but with the suffix _ms.
Configuring the Sybase Monitor on page 311
Sybase Performance Counters on page 314
Chapter 37 Sybase Database Server Monitoring
310
You need to set up the Sybase monitor environment before you can
configure the monitor.
To set up the Sybase monitor environment:
1 Install the Sybase client files and libraries on the Controller machine.
2 Verify a connection between the client and server on the Controller
machine. To do so, use the Sybase clients dsedit tool to ping the Adaptive
Server Enterprise Monitor Server.
Note: The port you use to monitor a Sybase server through a firewall
depends on the configuration of the Sybase server. Configuration
information for the connection between the client and server is located in
the Sybase client sql.ini file.
311
Configuring the Sybase Monitor
After you have set up the Sybase monitor environment, you can select the
objects you want the monitor to measure.
To configure the Sybase monitor:
Profile.
Click Monitors.
2 Under Database Server Resource Graphs, select Sybase (SiteScope) and
312
4 Click Next. The Configuring Sybase monitor page opens.
Enter the following server information:
Server. Type the name of the server (usually the same name as the Sybase
server but with the suffix _ms).
Username. Type the login name for the Sybase server.
Password. Type the password for the Sybase server.
Update every. The number entered here indicates the amount of time, in
seconds, between one monitor check and the next. By default the
monitor updates every 30 seconds.
313
For a list of the available performance counters, see Sybase Performance
profile.
Note: Certain measurements or counters are especially useful for
determining server performance and isolating the cause of a bottleneck
during an initial stress test on the SQL Server. For more information about
these counters, see the section that describes useful counters for stress
testing in the HP LoadRunner Analysis User Guide.
314
Sybase Performance Counters
The following measurements are available when monitoring a Sybase server:
Network Average packet size (Read) Reports the number of network
packets received
Average packet size (Send) Reports the number of network
packets sent
Network bytes (Read) Reports the number of bytes
received, over the sampling interval
Network bytes (Read)/sec Reports the number of bytes
received, per second
Network bytes (Send) Reports the number of bytes sent,
over the sampling interval
Network bytes (Send)/sec Reports the number of bytes sent,
per second
Network packets (Read) Reports the number of network
packets received, over the sampling
interval
Network packets (Read)/sec Reports the number of network
packets received, per second
Network packets (Send) Reports the number of network
packets sent, over the sampling
interval
Network packets (Send)/sec Reports the number of network
packets sent, per second
Memory Memory Reports the amount of memory, in
bytes, allocated for the page cache
315
Disk Reads Reports the number of reads made
from a database device
Writes Reports the number of writes made
to a database device
Waits Reports the number of times that
access to a device had to wait
Disk Grants Reports the number of times access
to a device was granted
Engine Server is busy (%) Reports the percentage of time
during which the Adaptive Server is
in a "busy" state
CPU time Reports how much "busy" time was
used by the engine
Logical pages (Read) Reports the number of data page
reads, whether satisfied from cache
or from a database device
Pages from disk (Read) Reports the number of data page
reads that could not be satisfied
from the data cache
Pages stored Reports the number of data pages
written to a database device
Stored
Procedures
Executed (sampling period) Reports the number of times a
stored procedure was executed, over
the sampling interval
Executed (session) Reports the number of times a
stored procedure was executed,
during the session
Average duration (sampling
period)
Reports the time, in seconds, spent
executing a stored procedure, over
the sampling interval
Average duration (session) Reports the time, in seconds, spent
executing a stored procedure,
during the session
316
Locks % Requests Reports the percentage of successful
requests for locks
Locks count Reports the number of locks. This is
an accumulated value.
Locks Granted immediately Reports the number of locks that
were granted immediately, without
having to wait for another lock to
be released
Granted after wait Reports the number of locks that
were granted after waiting for
another lock to be released
Not granted Reports the number of locks that
were requested but not granted
Wait time (avg.) Reports the average wait time for a
lock
SqlSrvr Locks/sec Reports the number of locks. This is
an accumulated value.
% Processor time (server) Reports the percentage of time that
the Adaptive Server is in a "busy"
state
Transactions Reports the number of committed
Transact-SQL statement blocks
(transactions)
Deadlocks Reports the number of deadlocks
317
Cache % Hits Reports the percentage of times that
a data page read could be satisfied
from cache without requiring a
physical page read
Pages (Read) Reports the number of data page
or from a database device
Pages (Read)/sec Reports the number of data page
or from a database device, per
second
Cache Pages from disk (Read) Reports the number of data page
reads that could not be satisfied
from the data cache
Pages from disk (Read)/sec Reports the number of data page
reads, per second, that could not be
satisfied from the data cache
Pages (Write) Reports the number of data pages
written to a database device
Pages (Write)/sec Reports the number of data pages
written to a database device, per
second
318
Process % Processor time (process) Reports the percentage of time that
a process running a given
application was in the "Running"
state (out of the time that all
processes were in the "Running"
state)
Locks/sec Reports the number of locks, by
process. This is an accumulated
value.
% Cache hit Reports the percentage of times that
a data page read could be satisfied
from cache without requiring a
physical page read, by process
Pages (Write) Reports the number of data pages
written to a database device, by
process
Transaction Transactions Reports the number of committed
(transactions), during the session
Rows (Deleted) Reports the number of rows deleted
from database tables during the
session
319
Transaction Inserts Reports the number of insertions
into a database table during the
session
Updates Reports the updates to database
tables during the session
Updates in place Reports the sum of expensive, in-
place and not-in-place updates
(everything except updates
deferred) during the session
Transactions/sec Reports the number of committed
(transactions) per second
Rows (Deleted)/sec Reports the number of rows deleted
from database tables, per second
Inserts/sec Reports the number of insertions
into a database table, per second
Updates/sec Reports the updates to database
tables, per second
Updates in place/sec Reports the sum of expensive, in-
place and not-in-place updates
(everything except updates
deferred), per second
320
Part X
J2EE Performance Monitoring
322
323
38
J2EE Performance Monitoring
The J2EE performance monitor provides complete insight into the J2EE
components on the application server (Servlets, JSPs, EJBs, JNDI, JDBC, and
DB SQL calls).
About J2EE Performance Monitoring on page 323
Setting Up the Monitoring Enviroment on page 324
Configuring the J2EE Monitor on page 329
J2EE Performance Counters on page 331
Examples of Modifying Application Server Configuration on page 332
About J2EE Performance Monitoring
The J2EE monitor provides the following information for each J2EE
component:
Average response time per method/query
Number of method calls per second
With such coverage of the J2EE architecture, users can get an overview of
the entire activity within the system. They can very easily correlate the end
user response time with the Web server activity (Servlets and JSPs data),
application server activity (JNDI and EJB's), and back-end activity of
database requests (JDBC methods and SQL queries).
Chapter 38 J2EE Performance Monitoring
324
The J2EE Monitor allows Performance Center users to analyze J2EE
component metrics during a load test run by using an agent which is
installed on the application server to collect information on the J2EE
components. These measurements are sent from the application server back
to the Performance Center Host through a Web server contained in the J2EE
monitor. The J2EE Monitor supports the leading applications servers, such
as: IBM WebSphere, BEA WebLogic, Oracle 9iAS and JBoss. For information
about the supported application servers, see Support Matrix on page 324.
Note: The J2EE Monitor requires MSXML 3.0 and later (this is included in
Internet Explorer 6.0). You can install MSXML 3.0 from the Microsoft MSDN
Web site (http://msdn2.microsoft.com/en-us/xml/bb190622.aspx).
Setting Up the Monitoring Enviroment
To monitor J2EE objects, you must first install and activate the J2EE monitor
on the application server machine. You then configure the J2EE monitor on
the client machine by selecting the counters you want the monitor to
measure.
You can monitor Java 2 Platform, Enterprise Edition (J2EE) objects on a
WebLogic, WebSphere, Oracle 9iAS, or JBoss application server during a load
test run using the J2EE performance monitor.
Installing the J2EE Monitor on the Application Server
To monitor J2EE objects, you must first install and activate the J2EE monitor
on the application server machine.
Support Matrix
Application Server Version Platform
WebLogic 4.x; 5.x; 6.x; 7.0; 8.1 Windows; Solaris; AIX
WebSphere 3.x; 4.x Windows; Solaris; AIX
325
To install the J2EE monitor on the application server:
1 Create a home directory on the application server machinefor example,
J2EEMonitor, and unzip the installation file <Performance Center
CD>\Additional Components\J2EE_Monitors\jmonitor_<platform>.jar file
into that directory.
If you do not have WinZip to unzip the installation file, use the following
command line to extract the installation file:
UNIX scripts extracted from the jar file may lose their execute permissions.
To fix this, open the J2EEMonitor Home Directory, and change the
permissions using the command line: chmod +x *.sh.
2 Open the J2EE Monitor Initializer by double-clicking <J2EEMonitor Home
Directory>\lib\sipatool.jar.
On a UNIX platform, or if the .jar extension in your system is not associated
with the Java run time environment, run sipatool.jar as follows:
Oracle 9iAS 1.0.2.2 Windows; Solaris; AIX
JBoss 2.4.x; 3.04 Windows; Solaris; AIX
<JDK>\bin\jar.exe -xf <installation file>
Application Server Version Platform
326
From the <J2EEMonitor Home Directory>\lib directory, type
Note: If the J2EE Monitor Initializer is not displayed, run sipatool.jar as
follows:
From the <J2EEMonitor Home Directory>\lib directory, type
<Java path>\ -jar sipatool.jar -nogui. Use the full path of the Java process of
the application server JDK.
3 In the J2EE Monitor Initializer dialog box, type the path to the application
server Java home directory, and click OK to run the tool.
4 Add -Xbootclasspath/p:<J2EEMonitor Home Directory>\classes\boot to the
application server command line arguments.
For more information about syntax for WebLogic, WebSphere, Oracle 9iAS,
or JBoss application servers, see Examples of Modifying Application Server
Configuration on page 332.
Initial J2EE Monitor Configuration Settings
During installation, the following was configured automatically on the J2EE
monitor application server:
Hooking mechanism. The J2EE monitor uses the J2EE Monitor Initializer
and Java hooking library.
Operation mode. The J2EE monitor uses the Auto Discovery operating
mode. In this mode, the system automatically discovers the J2EE
components (Servlet, JSP, JNDI, EJB and JDBC) that actually participate in
the business process.
JDBC information retrieval. The JDBC information retrieval setting
determines which data to return from the JDBC call. By default, the J2EE
monitor aggregates the measured data according to the JDBC operation,
for example: SELECT,UPDATE,CREATE. To modify this configuration, see
Configuring JDBC Information Retrieval on page 327.
java -jar sipatool.jar
327
EJB information retrieval. The EJB information retrieval setting
determines which data to return from the EJB call. By default, the J2EE
monitor is not configured to measure container methods, (for example,
ejbPassivate(), ejbCreate()). To modify this configuration, see
Configuring the EJB Information Retrieval on page 328.
Note: For information about alternative configuration settings, contact the
HP Software Support Web site (http://www.hp.com/go/hpsoftwaresupport).
Configuring JDBC Information Retrieval
To configure JDBC information retrieval:
1 Open <J2EEMonitor Home Directory>\etc\dispatcher.properties.
2 In the property sql.parsing.mode, type one of the following:
To measure the JDBC the method calls, like any other (non-JDBC)
measured method calls, type 1.
To aggregate the measured data according to the JDBC operation, for
example: SELECT, UPDATE, CREATE, type 2.
To aggregate the measured data according to specific SQL statement
(including the operation, the tables it acted on, and other parameters of
this statement), type 3.
Note: SQL Statements that exceed 3000 characters in length are not
supported.
328
Configuring the EJB Information Retrieval
To configure EJB information retrieval to include container methods:
1 Open <J2EEMonitor Home Directory>\dat\java_monitor.ini.
2 In the EJB_CONFIG section of the file, change the
setting to the following:
Starting the Auto Discovery Process
In Auto Discovery mode (the J2EE monitors default operating mode), the
system discovers which methods of the components (Servlet, JSP, JNDI, EJB
and JDBC) are participating in your business process and measures those
objects only.
Before configuring the J2EE monitor you need to start the Auto Discovery
process. To do this, start the application server, and run the Vuser script that
you intend to use in your load test against the application server. This
provides the Performance Center Host with a list of measurements that will
be available for monitoring.
Note: The next time you run the same script, you do not need to run a Vuser
before selecting the methods and counters you want to monitor.
hook_files=auto_detect
hook_files=auto_detect_container
329
Configuring the J2EE Monitor
To monitor J2EE performance, you must select the counters you want the
J2EE monitor to measure. You select these counters using the J2EE Choose
Measurements page.
To configure the J2EE monitor:
Profile.
Click Monitors.
2 Under Java Performance Graphs, select J2EE and click Next. The Choose
Server page opens.
monitor.
330
5 Select the methods and counters that you want to monitor.
For a list of the available performance counters, see J2EE Performance
profile.
331
Note: The size of a measurement name that can be displayed in the Analysis
is limited to 255 characters. If a measurement name exceeds this limit, the
counter name is truncated, and given a unique ID (UID). If you monitor
different events or make cross result graphs on the same counter, the UID
will remain the same.
The measurement name is truncated as follows:
standard prefix/counter truncated name<UID>/monitored event
For example:
/DB/JDBC/weblogic.jdbc.rmi.SerialPreparedStatement/int
executeUpdate()/INSERT INTO orders ( orderid _ userid _ orderdate _
shipaddr1 _ shipaddr2 _ shipcity _ shipstate _ shipzip _ shipcountry _
billaddr1 _ billaddr2 _ b <1> /Average Response Time
The full measurement name appears in the Measurement Description box.
J2EE Performance Counters
The following counters can be monitored for each method:
Average Response Time The average response time, in milliseconds, of the
J2EE object being monitored.
Method Calls per Second The number of J2EE object method calls per
second.
332
Examples of Modifying Application Server Configuration
When you installed HPs J2EE monitor files on your application server, you
configured it to run with J2EE monitor support. This section provides
examples for modifying the configuration of the following application
servers:
WebLogic - Version 4.x-5.x
WebLogic - Version 6.x
WebLogic - Version 8.1
WebSphere Server - Version 3.x
Oracle 9iAS Server
JBoss 2.4.x-3.04 Server
Note: It is important to set the environment variables in the order in which
they appear below.
333
WebLogic - Version 4.x-5.x
To configure the WebLogic 4.x-5.x server:
1 Copy the <WebLogic Home>\startWeblogic.cmd file into <WebLogic
Home>\startWeblogicMercury.cmd so that the file is backed up.
2 Open the <WebLogic Home>\startWeblogicMercury.cmd file.
3 Just before the Java command line used to start the server, add the following
variables:
For Windows platforms:
For UNIX platforms (csh):
4 In the same section of the file, add the following parameter to the Java
command line:
For example:
5 Run the <WebLogic Home>\startWeblogicMercury.cmd file.
set MERC_MONITOR_HOME=<J2EEMonitor Home Directory>
set JAVA_CLASSPATH=%JAVA_CLASSPATH%;
%MERC_MONITOR_HOME%\dat;
%MERC_MONITOR_HOME%\classes\xerces.jar
MERC_MONITOR_HOME <J2EEMonitor Home Directory>
JAVACLASSPATH=$JAVACLASSPATH:
$MERC_MONITOR_HOME/classes/xerces.jar
-Xbootclasspath/p:%MERC_MONITOR_HOME%\classes\boot
%JAVA_HOME%\bin\java -ms64m -mx64m -
-Dweblogic.class.path=%WEBLOGIC_CLASSPATH% -Dweblogic.home=.
-Djava.security.manager
-Djava.security.policy==.\weblogic.policy weblogic.Server
334
To configure the WebLogic 6.x server:
1 Copy the <WebLogic Home>\config\<domain name>\startWeblogic.cmd
file into <WebLogic Home>\config\<domain
name>\startWeblogicMercury.cmd so that the file is backed up.
2 Open the <WebLogic Home>\config\<domain name>\
startWeblogicMercury.cmd file.
3 Just before the java command line used to start the server, add the following
variables:
For UNIX platforms:
4 In the same section of the file add a parameter to the command line:
For example:
5 Run the <WebLogic Home>\config\<domain name>\
set CLASSPATH=%CLASSPATH%;%MERC_MONITOR_HOME%\dat;
MERC_MONITOR_HOME=<J2EEMonitor Home Directory>
CLASSPATH=$CLASSPATH:$MERC_MONITOR_HOME/dat:
"%JAVA_HOME%\bin\java" -hotspot -ms64m -mx64m
-classpath %CLASSPATH% -Dweblogic.Domain=mydomain
-Dweblogic.Name=myserver "-Dbea.home=f:\bea" "
-Djava.security.policy==f:\bea\wlserver6.0/lib/weblogic.policy"
-Dweblogic.management.password=%WLS_PW% weblogic.Server
335
To configure the WebLogic 7.x server:
1 Copy the <WebLogic Home>\server\bin\startwls.cmd file into <WebLogic
Home>\server\bin\startwlsMercury.cmd so that the file is backed up.
2 Open the <WebLogic Home>\server\bin\startwlsMercury.cmd file.
variables:
For UNIX platforms:
4 In the same section of the file add a parameter to the command line:
For example:
5 Copy the <domain name>\startWeblogic.cmd file into <domain
6 Open the <domain name>\startWeblogicMercury.cmd file.
336
7 Find the call to the Weblogic server. For example:
8 Change the call from startWLS.cmd to startWLSMercury.cmd, and save the
file.
WebLogic - Version 8.1
To configure the WebLogic 8.1 server:
1 Copy the <WebLogic Home>\samples\domains\<domain
name>\startWeblogic.cmd file into <WebLogic
Home>\samples\domains\<domain name>\startWeblogicMercury.cmd so
that the file is backed up.
2 Open the <WebLogic Home>\samples\domains\<domain name>\
variables:
For UNIX platforms:
In the same section of the file add a parameter to the command line:
call D:\bea\weblogic700\server\bin\startWLS.cmd
337
For example:
4 Copy the <domain name>\startWeblogic.cmd file into <domain
5 Open the <domain name>\startWeblogicMercury.cmd file.
6 Find the call to the Weblogic server. For example:
7 Change the call from startWeblogic.cmd to startWeblogicMercury, and save
the file.
Note: WebLogic 8.1 supports SUN JVM only. JRockit JVM is available with
WebLogic 8.1 using the J2EE Diagnostics monitor only.
call D:\bea\weblogic8\server\bin\startWLS.cmd
338
By default, the WebSphere 3.x application server runs on Windows as an
automatic service, upon machine startup. Since HP does not currently
support Performance Center J2EE monitoring on a WebSphere server run as
an automatic service, you must change the default WebSphere server startup
to manual.
To change the default WebSphere 3.x server startup:
1 Select Start > Settings > Control Panel.
2 Double-click Services.
3 Select IBM WS AdminServer, and click the Stop button.
4 Double-click IBM WS AdminServer, and select the Manual Startup Type.
5 Click OK to save your settings and close the dialog box.
You can now start the WebSphere Server from <WebSphere
Home>\AppServer\bin\debug\adminserver.bat, instead of using the
automatic service.
To add Performance Center J2EE monitor support to the WebSphere 3.x
server:
1 Make a backup copy of <WebSphere Home>\AppServer\bin\debug\
adminserver.bat.
2 Open <WebSphere Home>\AppServer\bin\debug\adminserver.bat.
3 Add the following environment variables at the end of the SET_CP section:
For UNIX platforms:
export MERC_MONITOR_HOME
339
4 Run the adminserver.bat file.
5 Open the WebSphere Advanced Administrative Console, and select
View > Topology.
6 Expand the WebSphere Administrative Domain tree by selecting
<server machine name> > Default Server.
7 Select the General tab in the Application Server:Default Server window.
8 Add
to the command line Arguments box, and click Apply.
9 If you are working with a WebSphere 3.0 Server with JDK1.1.7 IBM, double-
click Environment.
a In the Variable Name box, type _CLASSLOAD_HOOK.
b In the Value box, type jdkhook.
Click Add, then OK, and then Apply.
10 For Windows 2000/NT or Solaris, from the General tab, open the
Environment Editor dialog box and add the following variables to the
Environment box:
For Windows 2000/NT:
For Solaris:
Click OK to close the Environment Editor dialog box.
11 Close the WebSphere Advanced Administrative Console.
12 Close and restart the adminserver.bat file.
name=CLASSPATH
value=<J2EEMonitor Home Directory>\dat
name=CLASSPATH
value=<J2EEMonitor Home Directory>/dat
340
You can start the WebSphere 4.x server using the startServerBasic.bat file or
the startServer.bat file.
To configure the WebSphere 4.x server:
1 Check that the WebSphere Administrative Server is running, and start the
Administrator Console.
2 In the WebSphere Administrative Domain tree, expand the Nodes,
Hostname, and Application Servers subtrees, and select the Default Server
(or the application server you wish to use with J2EE monitor).
3 Right-click the Default Server, select Properties from the menu, and click the
General tab.
4 For Windows 2000/NT or Solaris, from the General tab, open the
Environment Editor dialog box and add the following variables to the
Environment box:
For Windows 2000/NT:
For Solaris:
Click OK to close the Environment Editor dialog box.
5 Click the Advanced JVM Settings tab and select Advanced JVM settings. In
the Command line arguments field, add the following value for Windows
2000/NT, Solaris, and AIX:
6 Click OK and Apply to save the changes for the application server. You can
now start and stop your WebSphere server using the Performance Center
J2EE monitor.
name=CLASSPATH
value=<J2EEMonitor Home Directory>\dat;
<J2EEMonitor Home Directory>\classes\xerces.jar
name=CLASSPATH
value=<J2EEMonitor Home Directory>/dat:
<J2EEMonitor Home Directory>/classes/xerces.jar
341
Oracle 9iAS Server
1 Edit the file env.cmd (env.sh in UNIX platforms) as follows:
The JAVA_HOME environment variable should point to the location of
the Java Virtual machine used to run the application server.
The DETECTOR_INS_DIR environment variable should point to the
location of the monitor installation.
The APP_SERVER_DRIVE environment variable should specify the drive
hosting the application server installation (for example, D:). Do not
modify this variable on UNIX platforms.
The APP_SERVER_ROOT environment variable should specify the
application server root directory.
2 Run the oc4jMonitor.cmd (oc4jMonitor.sh on UNIX platforms).
JBoss 2.4.x-3.04 Server
1 Make a backup copy of <JBoss Home>\run.bat (run.sh on UNIX platforms)
file into <JBoss Home>\runMercury.bat (runMercury.sh for UNIX).
2 Open <JBoss Home>\runMercury.bat (runMercury.sh on UNIX).
Just before the Java command line used to start the server, add the following
variables:
For UNIX platforms:
342
3 In the same section of the file, add the following parameter to the command
line:
For example:
4 Run the <JBoss Home>\runMercury.bat file (runMercury.sh on UNIX
platforms).
%JAVA_HOME%\bin\java -ms64m -mx64m -
Xbootclasspath/p:%MERC_MONITOR_HOME%\classes\boot
-Dweblogic.class.path=%WEBLOGIC_CLASSPATH% -Dweblogic.home=.
-Djava.security.manager
-Djava.security.policy==.\weblogic.policy weblogic.Server
Part XI
Application Deployment Monitoring
344
345
39
Application Deployment Solutions
Using Performance Centers Application Deployment Solution monitor, you
can isolate server performance bottlenecks by monitoring the Citrix
MetaFrame XP or 1.8 server during a load test run.
Configuring the Citrix MetaFrame Server Monitor on page 347
Citrix MetaFrame Performance Counters on page 350
Chapter 39 Application Deployment Solutions
346
The Citrix MetaFrame XP monitor displays statistics about the resource
usage on the Citrix MetaFrame XP or 1.8 server machine during the load test
run.
To monitor the Citrix server performance, you must first activate the Citrix
MetaFrame XP monitor on the application server machine and enable the
counters you want to monitor on the Citrix server.
Note: The port you use to monitor a Citrix MetaFrame server through a
firewall depends on the configuration of your server.
Before configuring the monitor:
1 From the Performance Center Host machine, map a network drive to the
Citrix server machine. This provides the required authentication to the
Performance Center Host to access the resource counters.
2 Launch PerfMon from the Performance Center Host machine to enable the
counters on the Citrix server. This allows you to monitor the same counters
for the ICA Session object on the Citrix monitor.
3 To provide the Performance Center Host with a list of measurements that
will be available for monitoring, you must first initialize Vusers before
running the load test. After you have initialized the Vusers, you can
configure the Citrix Monitor and add the ICA Session counters.
4 Although SiteScope is installed on the Utility Server, we recommend
347
Note:
Measurements that monitor instances are valid for the currently running
Citrix session only. If you run this load test again, you will need to
reconfigure the measurements that are instance-oriented.
To monitor the different instances, check that the server login and logout
procedures are recorded in the Vuser_init and Vuser_end sections
respectively, and not in the Action section of the script. For more
information, see the HP Virtual User Generator User Guide.
Configuring the Citrix MetaFrame Server Monitor
After you have set up the monitoring enviroment, you select the counters
you want the Citrix MetaFrame XP monitor to measure. You select these
counters using the Citrix MetaFrame XP Choose Measurements page.
To configure the Citrix MetaFrame Server monitor:
Profile.
Click Monitors.
348
2 Under Application Deployment Solutions Graphs, select Citrix MetaFrame
XP or Citrix MetaFrame XP (SiteScope) and click Next. The Choose Server
page opens.
For Citrix MetaFrame XP Server, type the user login name and password
on the Citrix server, and click Next.
For Citrix MetaFrame XP Server (SiteScope), type the SiteScope machine
name and port (default: 8888), and specify if you are using a Secure HTTP
4 Click Next. The Add Monitor page opens. Verify the server and the update
frequency, and click Next.

349
For a list of available performance counters, see Citrix MetaFrame
profile.
Note: If the dialog box freezes after clicking Add, you may need to rebuild
the localhost cache on the Citrix server machine. For more information, see
Document IDs CTX003648 and CTX759510 in the Citrix Knowledge Base
(http://knowledgebase.citrix.com/cgi-bin/webcgi.exe?New,KB=CitrixKB).
350
Citrix MetaFrame Performance Counters
The following tables describe some of the counters that can be measured.
Non-Virtual Counters
% Disk Time The percentage of elapsed time that the selected
disk drive is busy servicing read or write requests.
% Processor Time The percentage of time that the processor is
executing a non-Idle thread. This counter is a
primary indicator of processor activity. It is
calculated by measuring the time that the processor
spends executing the thread of the Idle process in
each sample interval, and subtracting that value
from 100%. (Each processor has an Idle thread
which consumes cycles when no other threads are
ready to run). It can be viewed as the percentage of
the sample interval spent doing useful work. This
counter displays the average percentage of busy
time observed during the sample interval. It is
calculated by monitoring the time the service was
inactive, then subtracting that value from 100%.
File data Operations/sec The rate that the computer is issuing Read and
Write operations to file system devices. It does not
include File Control Operations.
351
Interrupts/sec The average number of hardware interrupts the
processor is receiving and servicing in each second.
It does not include DPCs, which are counted
separately. This value is an indirect indicator of the
activity of devices that generate interrupts, such as
the system clock, the mouse, disk drivers, data
communication lines, network interface cards and
other peripheral devices. These devices normally
interrupt the processor when they have completed a
task or require attention. Normal thread execution
is suspended during interrupts. Most system clocks
interrupt the processor every 10 milliseconds,
creating a background of interrupt activity. This
counter displays the difference between the values
observed in the last two samples, divided by the
duration of the sample interval.
Output Session Line Speed This value represents the line speed from server to
client for a session in bps.
Input Session Line Speed This value represents the line speed from client to
server for a session in bps.
Page Faults/sec A count of the Page Faults in the processor. A page
fault occurs when a process refers to a virtual
memory page that is not in its Working Set in main
memory. A Page Fault will not cause the page to be
fetched from disk if that page is on the standby list,
and hence already in main memory, or if it is in use
by another process with whom the page is shared.
352
Pages/sec The number of pages read from the disk or written
to the disk to resolve memory references to pages
that were not in memory at the time of the
reference. This is the sum of Pages Input/sec and
Pages Output/sec. This counter includes paging
traffic on behalf of the system Cache to access file
data for applications. This value also includes the
pages to/from non-cached mapped memory files.
This is the primary counter to observe if you are
concerned about excessive memory pressure (that
is, thrashing), and the excessive paging that may
result.
Pool Nonpaged Bytes The number of bytes in the Nonpaged Pool, a
system memory area where space is acquired by
operating system components as they accomplish
their appointed tasks. Nonpaged Pool pages cannot
be paged out to the paging file, but instead remain
in main memory as long as they are allocated.
Private Bytes The current number of bytes this process has
allocated that cannot be shared with other
processes.
Processor Queue Length The instantaneous length of the processor queue in
units of threads. This counter is always 0 unless you
are also monitoring a thread counter. All processors
use a single queue in which threads wait for
processor cycles. This length does not include the
threads that are currently executing. A sustained
processor queue length greater than two generally
indicates processor congestion. This is an
instantaneous count, not an average over the time
interval.
Threads The number of threads in the computer at the time
of data collection. Notice that this is an
instantaneous count, not an average over the time
interval. A thread is the basic executable entity that
can execute instructions in a processor.
Latency Session Average The average client latency over the life of a session.
353
Virtual Channel Counters
Latency Last Recorded Represents the last recorded latency measurement
for this session.
Latency Session
Deviation
The difference between the minimum and
maximum measured values for a session.
Input Session Bandwidth The bandwidth from client to server traffic for a
session in bps.
Input Session
Compression
The compression ratio for client to server traffic for
a session.
Output Session Bandwidth The bandwidth from server to client traffic for a
session (in bps).
Output Session
Compression
The compression ratio for server to client traffic for
a session.
Output Session Linespeed The line speed from server to client for a session (in
bps).
Input Audio Bandwidth The bandwidth from client to server traffic on the
audio mapping channel (in bps).
Input Clipboard
Bandwidth
The bandwidth from client to server traffic on the
clipboard mapping channel (in bps)
Input COM1 Bandwidth The bandwidth from client to server traffic on the
COM1 channel (in bps).
Input COM2 Bandwidth The bandwidth from client to server traffic on the
Input COM Bandwidth The bandwidth from client to server traffic on the
COM channel (in bps).
Input Control Channel
Bandwidth
ICA control channel (in bps).
354
Input Drive Bandwidth The bandwidth from client to server traffic on the
client drive mapping channel (in bps).
Input Font Data
Bandwidth
local text echo font and keyboard layout channel
(in bps).
Input Licensing
Bandwidth
The bandwidth from server to client traffic on the
licensing channel (in bps).
Input LPT1 Bandwidth The bandwidth from client to server traffic on the
LPT1 channel (in bps).
Input LPT2 Bandwidth The bandwidth from client to server traffic on the
Input Management
Bandwidth
client management channel (in bps).
Input PN Bandwidth The bandwidth from client to server traffic on the
Program Neighborhood channel (in bps).
Input Printer Bandwidth The bandwidth from client to server traffic on the
printer spooler channel (in bps).
Input Seamless Bandwidth The bandwidth from client to server traffic on the
Seamless channel (in bps).
Input Text Echo
Bandwidth
local text echo data channel (in bps).
Input Thinwire Bandwidth The bandwidth from client to server traffic on the
Thinwire (graphics) channel (in bps).
Input VideoFrame
Bandwidth
Tthe bandwidth from client to server traffic on the
VideoFrame channel (in bps).
Output Audio Bandwidth The bandwidth from server to client traffic on the
audio mapping channel (in bps).
Output Clipboard
Bandwidth
clipboard mapping channel (in bps).
Output COM1 Bandwidth The bandwidth from server to client traffic on the
355
Output COM2 Bandwidth The bandwidth from server to client traffic on the
Output COM Bandwidth The bandwidth from server to client traffic on the
COM channel (in bps).
Output Control Channel
Bandwidth
ICA control channel (in bps).
Output Drive Bandwidth The bandwidth from server to client traffic on the
client drive channel (in bps).
Output Font Data
Bandwidth
local text echo font and keyboard layout channel
(in bps).
Output Licensing
Bandwidth
licensing channel (in bps).
Output LPT1 Bandwidth The bandwidth from server to client traffic on the
Output LPT2 Bandwidth The bandwidth from server to client traffic on the
Output Management
Bandwidth
client management channel (in bps).
Output PN Bandwidth The bandwidth from server to client traffic on the
Program Neighborhood channel (in bps).
Output Printer Bandwidth The bandwidth from server to client traffic on the
printer spooler channel (in bps).
Output Seamless
Bandwidth
Seamless channel (in bps).
Output Text Echo
Bandwidth
local text echo data channel (in bps).
Output Thinwire
Bandwidth
Thinwire (graphics) channel (in bps).
Output VideoFrame
Bandwidth
VideoFrame channel (in bps).
356
Part XII
ERP/CRM Server Resource Monitoring
358
359
40
Introducing ERP/CRM Server Resource
Monitoring
The ERP/CRM server resource monitors provide you with performance
information for SAP, SAPGUI, SAP Portal, Siebel Server Manager, Siebel Web
Server, and PeopleSoft (Tuxedo) servers. To monitor data for SAP CCMS
(Computer Center Management System), use the SiteScope monitor. For
more information, see Configuring the SiteScope Monitor on page 92.
To obtain this data, you need to activate the ERP/CRM server resource
monitors before running the load test, and indicate which statistics and
measurements you want to monitor.
Note: Performance Center no longer provides the SAP monitor. To monitor
SAP server resources, use the SAPGUI monitor. Performance Center supports
SAP monitor profiles for load tests saved in previous versions of Performance
Center.
Chapter 40 Introducing ERP/CRM Server Resource Monitoring
360
361
41
SAPGUI Monitoring
The SAPGUI monitor displays statistics about the resource usage of a SAP R/3
system during the load test run.
You can use the SAPGUI monitor to view:
the number of configured instances for each SAP system
data for all application instances (not just the one you logged on to)
transactions used and the users that call them
number of users working on the different instances
performance history for recent periods of all instances
response time distribution
resource consumption for any application server
application server workload for today or for a recent period
Configuring the SAPGUI Monitor on page 362
SAPGUI Performance Counters on page 365
Chapter 41 SAPGUI Monitoring
362
Before monitoring a SAP R/3 system server, you must enable the SAPGUI
monitor.
To enable the SAPGUI monitor:
1 Install the SAPGUI for Windows 6.20 client on the Performance Center Host
machine.
2 Install the latest patch for the SAPGUI for Windows 6.20 client. The lowest
supported level is patch 36. (SAPGUI Patches can be downloaded from
https://websmp104.sap-ag.de/patches.)
3 Check that you can access the st03 transaction and query for "last minute
load" information with the user name and password defined in the
Performance Center host.
Note: To monitor SAPGUI, you must change the DCOM configuration of
Wlrun.Engine on the Controller machine to work as an Interactive User. To
do so:
1. Make sure wlrun process is not running on the machine.
2. Select Run from the Windows Start menu and type dcomcnfg.
3. Press ENTER to display the list of DCOM applications.
4. In the displayed application list, double-click wlrun.LrEngine.
5. In the Identity tab of the dialog box, select The interactive user.
6. Click OK twice to close both dialog boxes.
Configuring the SAPGUI Monitor
To obtain data from the SAPGUI monitor, you need to indicate which
statistics and measurements you want to monitor.
To configure the SAPGUI monitor:
363
Profile.
Click Monitors.
2 Under ERP/CRM Server Resource Graphs, select SAPGUI and click Next. The
3 Type the following information:
Server Name. Type the name or IP address of the server whose resources
User Name. Type the login name used to access the SAPGUI server.
Password. Type the password for the login name.
SAP Server. Type the name of the SAPGUI server.
Client. Type the number used in the Client field of the SAP logon details.
You can type the server name in the following ways:
364
Type the server description, as it is displayed in the Server Name box.
Type a string, in the format server_network_name[:system_number],
where server_network_name is the name (or IP address) of the application
server as it is displayed in the Server Name field (for example:
pipeline.mercury.com). If a SAP router string is also specified, the
server_network_name should be the concatenation of the router string
and the application server.
365
instances of the highlighted measurement are running.
Click Add Measurement.
For a list of the available performance counters, see SAPGUI Performance
profile.
SAPGUI Performance Counters
The following table lists the most commonly monitored counters:
Average CPU time The average CPU time used in the work process.
Average response time The average response time, measured from the time
a dialog sends a request to the dispatcher work
process, through the processing of the dialog, until
the dialog is completed and the data is passed to the
presentation layer. The response time between the
SAP GUI and the dispatcher is not included in this
value.
Average wait time The average amount of time that an unprocessed
dialog step waits in the dispatcher queue for a free
work process. Under normal conditions, the
dispatcher work process should pass a dialog step to
the application process immediately after receiving
the request from the dialog step. Under these
conditions, the average wait time would be a few
milliseconds. A heavy load on the application server
or on the entire system causes queues at the
dispatcher queue.
Average load time The time needed to load and generate objects, such
as ABAP source code and screen information, from
the database.
366
Database calls The number of parsed requests sent to the database.
Database requests The number of logical ABAP requests for data in the
database. These requests are passed through the R/3
database interface and parsed into individual
database calls. The proportion of database calls to
database requests is important. If access to
information in a table is buffered in the SAP buffers,
database calls to the database server are not
required. Therefore, the ratio of calls/requests gives
an overall indication of the efficiency of table
buffering. A good ratio would be 1:10.
Roll ins The number of rolled-in user contexts.
Roll outs The number of rolled-out user contexts.
Roll in time The processing time for roll ins.
Roll out time The processing time for roll outs.
Roll wait time The queue time in the roll area. When synchronous
RFCs are called, the work process executes a roll out
and may have to wait for the end of the RFC in the
roll area, even if the dialog step is not yet
completed. In the roll area, RFC server programs can
also wait for other RFCs sent to them.
Average time per logical
DB call
The average response time for all commands sent to
the database system (in milliseconds). The time
depends on the CPU capacity of the database server,
the network, the buffering, and on the input/output
capabilities of the database server. Access times for
buffered tables are many magnitudes faster and are
not considered in the measurement.
367
42
SAP Portal Monitoring
The SAP Portal monitor displays statistics about the resource usage of a SAP
Portal system during the load test run.
Configuring the SAP Portal Monitor on page 367
SAP Portal Performance Counters on page 371
Configuring the SAP Portal Monitor
To obtain data on the SAP R/3 system server, you need to enable the SAP
online monitor (from the Performance Center Host) before executing the
load test, and indicate which statistics and measurements you want to
monitor.
To configure the SAP Portal monitor through the User Site:
Profile.
Chapter 42 SAP Portal Monitoring
368
Click Monitors.
2 Under ERP/CRM Server Resource Graphs, select SAP Portal (SiteScope) and
369
4 Click Next. The Configure Server page opens.
5 Type the following SAP Portal monitor information:
Application URL. Type the URL of the administrator server for the
application.
Username. Type the user name for the server administrator page.
Password. Type the password for the server administrator page.
HTTP Proxy. Type a proxy server to use, including the port (optional).
the user name.
Proxy Server Password. If the proxy server requires authorization, type
the password.
Update every. Type the amount of time lapse between the SiteScope
check of the monitor. The default is 30 seconds.
370
For a list of the available performance counters, see SAP Portal Performance
profile.
371
SAP Portal Performance Counters
The following table shows the default counters that can be measured:
Accumulated Amount of
Outbound Data (bytes)
The accumulated amount of outbound data,
measured in bytes.
Time for all Requests (ms) The total time, in milliseconds, taken for processing
all requests.
Average Amount of
Outbound Data per
Request (bytes)
The average amount of outbound data per request,
measured in bytes.
Average Number of
Component Calls per
Request (bytes)
The average number of component calls per request,
measured in bytes.
Average Time of a
Request (ms)
The average amount of time, in milliseconds, taken
to process a request.
Number of Calls with
Outbound Data
The total number of calls with outbound data.
Number of Component
Calls for all Requests
The total number of component calls for all requests.
Number of Requests since
First Request
The total number of requests since the first request
was made.
Requests per Second The number of requests made per second.
Time Stamp of First
Request
The time stamp of the first request.
372
373
43
Siebel Web Server Monitoring
The Siebel Web Server monitor displays statistics about the resource usage of
a Siebel Web Server during the load test run.
Configuring the Siebel Web Server Monitor on page 374
Siebel Web Server Performance Counters on page 378
Before monitoring a Siebel Web Server, you must set up the server monitor
environment.
To set up the Siebel Web Server monitor environment:
Check that the Siebel web server plug-in is installed and configured to
enable the display of the statistics you want to monitor. This may require
that stats page sections be enabled by editing the eapps.cfg file for the Siebel
server. For more information, see the Siebel documentation.
Chapter 43 Siebel Web Server Monitoring
374
monitoring. From the SiteScope Panel, select the monitor group polling the
Siebel Web Server machine, and check that the monitor displays a list of
server measurements in the Status column. This may require that stats page
sections be enabled by editing the eapps.cfg file for the Siebel server. Consult
the Siebel documentation for more information.
Configuring the Siebel Web Server Monitor
To obtain data on the Siebel Web Server, you need to enable the Siebel Web
Server online monitor (from the Performance Center Host) before running
the load test, and indicate which statistics and measurements you want to
monitor.
To configure the Siebel Web Server monitor through the User Site:
Profile.
Click Monitors.
375
2 Under ERP/CRM Server Resource Graphs, select Siebel Web Server
(SiteScope) and click Next. The Choose Server page opens.
376
4 Click Next. The Configure Server page opens.
5 Type the following Siebel Web Server monitor information:
Service. Type the name of a valid Siebel virtual directory (for example,
callcenter or sales), one whose URL is an entry point defined in the Siebel
SWSE configuration file (eapps.cfg).
Application URL. Displays the default URL of the Web plug-in server stats
page for the application you want to monitor. For example,
http://siebelsrv/<service>/_stats.swe. If the Siebel web server is
configured to support verbose mode, and you want to include
information on "Locks" and "Current Operations Processing," you can
add verbose=high, for example:
http://siebelsrv/<service>/_stats.swe?verbose=high.
Username. Type the user name to access the Web server stats page.
Password. Type the password to access the Web server stats page.
377
HTTP Proxy. If you are using a proxy to access the Siebel server, type the
proxy server to use including the port (for example,
proxy.sitescope.com:8080).
the proxy user name.
Proxy Server Password. If the proxy server requires authorization, type
the proxy password.
Update every. Type how frequently the monitor should check the
Database server.
For a list of the available performance counters, see Siebel Web Server
profile.
378
Siebel Web Server Performance Counters
Anonymous sessions requested
from the pool
The number of anonymous sessions
requested from the pool.
Open Session Time The time users experience logging
on to the system.
Anon Session Removed The number of anonymous sessions
removed from the pool.
Anon Session Available The number of anonymous sessions
available in the pool.
Anonymous sessions returns to
the pool
The number of anonymous sessions
returned to the pool.
Response Time The time taken to respond to a user
request.
Close Session Time The time users experience logging
off the system.
Request Time The time taken to process the user
request.
379
44
Siebel Server Manager Monitoring
The Siebel Server Manager monitor displays statistics about the resource
usage of a Siebel Server Manager during the load test run.
Setting up the Monitoring Environment on page 379
Configuring the Siebel Server Manager Monitor on page 381
Siebel Server Manager Performance Counters on page 382
Setting up the Monitoring Environment
Before you set up the monitor, perform the following:
Check that SiteScope (the application that is used to monitor the Siebel
Server Manager) has been installed on a server. You can install SiteScope on
the same machine as the Controller, or on a dedicated server.
On the machine where SiteScope is installed, configure SiteScope to monitor
the required Siebel Server Manager machine. For more information, see
Configuring the Siebel Server Manager Client on the SiteScope Machine
on page 380.
monitoring. From the SiteScope Panel, select the monitor group polling the
Siebel Server Manager machine, and check that the monitor displays a list of
server measurements in the Status column.
Chapter 44 Siebel Server Manager Monitoring
380
Configuring the Siebel Server Manager Client on the
SiteScope Machine
To monitor the Siebel Server Manager performance, you install the Siebel
Server Manager client on the SiteScope machine, and configure SiteScope to
monitor the required Siebel Server Manager machine.
To configure the Siebel Server Manager monitor on the SiteScope machine:
1 Restart SiteScope after installing the monitor add-in, and verify connectivity
to the Siebel SWSE page by opening the following URL from the machine
where sitescope is installed:
2 In the SiteScope main panel, select an existing group or create a new group.
3 Under Add to Group, click Monitor and select Siebel Server Manager from
the list of monitors.
4 Click Choose server, and type the name of the Siebel Server in the
Application server field, the Enterprise Server in the Enterprise server field,
and the Gateway Server in the Gateway server field.
5 If necessary, type the user name and password for the Siebel Server Manager
client.
6 Type the path to the Siebel Server Manager in the Path to Script field.
7 Click Browse, select your desired counters, and click Choose Counters.
8 In the Title field, type a name for the monitor.
Note: When you assign a name to a monitor, include the server name in the
monitor name. This avoids any confusion as to which host the monitor
belongs. For example, SiebelManager on sieb07.
Do not use "\" in the Title field.
9 Click Add Monitor.
http://<your_siebel_server>/callcenter/_stats.swe
381
Configuring the Siebel Server Manager Monitor
After installing and configuring the Siebel Server Manager client on the
SiteScope machine, you must select the counters to measure.
To configure the Siebel Server Manager monitor:
Profile.
Click Monitors.
2 Under ERP/CRM Server Resource Graphs, select Siebel Server Manager and
monitor.
382
For a list of the available performance counters, see Siebel Server Manager
profile.
Limitations
The SiteScope monitor has the following limitations when it retrieves
information from a Siebel Server Manager:
The component counters (for example, Average SQL Time for
<component>) are updated with the aggregated Siebel data only at the
end of a user session.
The SiteScope monitor consumes very high CPU resources
(approximately 40%).
Siebel Server Manager Performance Counters
Average Connect Time The average connection time.
Average Reply Size The average size of a user reply.
Average Request Size The average size of a user request.
Average Requests Per
Session
The average number of user requests per session.
Average Response Time The average amount of time that it takes the server to
respond to a request.
Average Think Time The average amount of think time taken to respond to
a request.
Avg SQL Execute Time The average SQL execute time.
Avg SQL Fetch Time The average SQL fetch time.
383
Avg SQL Parse Time The average SQL parse time.
CPU Time The CPU time used in the work process.
Elapsed Time The total amount of elapsed time.
Num of DBConn Retries The number of database connection retries.
Num of DLRbk Retries The number of DLRbk retries.
Num of Exhausted
Retries
The total number of retries that expired.
Number of SQL
Executes
The total number of SQL executes.
Number of SQL Fetches The total number of SQL fetches.
Number of SQL Parses The total number of SQL parses.
Number of Sleeps The number of sleeps.
Object Manager Errors The total number of object manager errors.
Reply Messages The total number of reply messages.
Request Messages The total number of request messages.
SQL Execute Time The total SQL execute time.
SQL Fetch Time The total SQL fetch time.
SQL Parse Time The total SQL parse time.
Sleep Time The total sleep time.
Tests Attempted The number of tests attempted.
Tests Failed The number of tests that failed.
Tests Successful The number of tests that were successful.
Total Reply Size The total reply size, measured in bytes.
Total Request Size The total request size, measured in bytes.
Total Response Time The total response time.
384
Total Tasks The total number of tasks.
Total Think Time The total think time.
385
45
PeopleSoft (Tuxedo) Monitoring
The PeopleSoft (Tuxedo) monitor displays statistics about the resource usage
of a PeopleSoft (Tuxedo) server during the load test run.
Setting Up the PeopleSoft (Tuxedo) Monitor on page 385
Configuring the PeopleSoft (Tuxedo) Monitor on page 386
PeopleSoft (Tuxedo) Performance Counters on page 389
Setting Up the PeopleSoft (Tuxedo) Monitor
If Tuxedo 7.1 or later is installed, you can monitor more than one PeopleSoft
(Tuxedo) application server at a time. If Tuxedo 6.5 or earlier is installed, you
can monitor only one PeopleSoft (Tuxedo) application server at a time.
Before you set up the monitor, perform the following:
Ensure that a Tuxedo workstation client (not a native client), version 6.3 or
later, is installed on the Controller machine. Use a Tuxedo 6.x client if a
Tuxedo 6.x server is used, and Tuxedo 7.1 or later client if a Tuxedo 7.1 or
later server is used. If you use a Tuxedo 6.5 or earlier server, you can still use
a Tuxedo 7.1 or later client in order to monitor it, provided that you set the
WSINTOPPRE71 environment variable to yes.
Chapter 45 PeopleSoft (Tuxedo) Monitoring
386
Note: A Tuxedo workstation client communicates with the application
server over the network, and is not required to run the Tuxedo
application server on the same machine. A native client can only
communicate with the Tuxedo application server if it is part of the
relevant Tuxedo domain.
Define the Tuxedo environment variables on the Performance Center
Host machineset the TUXDIR variable to the Tuxedo installation
directory, (for example, V:\environ\32\Tuxedo8.0), and add the Tuxedo bin
directory to the PATH variable.
Ensure that the workstation listener (WSL) process is running. This
enables the application server to accept requests from workstation
clients. The address and port number used to connect to the application
server must match those dedicated to the WSL process.
Note: For information on configuring the WSL, see the BEA Tuxedo Web
site (http://edocs.beasys.com/tuxedo/tux81/rf5/rf5101.htm#1534543).
Configuring the PeopleSoft (Tuxedo) Monitor
To obtain data for this graph, you need to enable the PeopleSoft (Tuxedo)
online monitor (from the Performance Center Host) and select the default
measurements you want to display, before running the load test.
To configure the PeopleSoft (Tuxedo) monitor through the User Site:
Profile.
387
Click Monitors.
2 Under ERP/CRM Server Resource Graphs, select PeopleSoft (Tuxedo) and
monitor.
Note: If you are using multiple instances of the Tuxedo monitor on the
same machine, then type the port number of each Tuxedo monitor in order
to distinguish one instance from another. The entry should have the
following format: <machine name>:<port number>.
Type the user name and password of the monitored server, and the name of
the client machine.
388
instances of the highlighted measurement are running.
Click Add Measurement.
profile.
For a list of the available performance counters, see PeopleSoft (Tuxedo)
profile.
389
PeopleSoft (Tuxedo) Performance Counters
The following table describes the default counters that can be measured. Pay
particular attention to the following measurements: % Busy Clients, Active
Clients, Busy Clients, Idle Clients, and all the queue counters for the
APPQ/PSAPPSRV queue.
Monitor Measurements
Machine % Busy Clients. The per cent of active clients
currently logged in to the Tuxedo application server
which are waiting for a response from the
application server.
Active Clients. The total number of active clients
currently logged in to the Tuxedo application server.
Busy Clients. The total number of active clients
that are waiting for a response from the application
server.
Current Accessers. Number of clients and servers
currently accessing the application either directly
on this machine or through a workstation handler
on this machine.
Current Transactions. Number of in use transaction
table entries on this machine.
Idle Clients. The total number of active clients
that are not waiting for a response from the
application server.
Workload Completed/second. The total workload
on all the servers for the machine that was
completed, per unit time.
Workload Initiated/second. The total workload on
all the servers for the machine that was initiated,
per unit time.
390
Queue % Busy Servers. The per cent of active servers
currently handling Tuxedo requests.
Active Servers. The total number of active servers
either handling or waiting to handle Tuxedo
requests.
Busy Servers. The total number of active servers
currently busy handling Tuxedo requests.
Idle Servers. The total number of active servers
currently waiting to handle Tuxedo requests.
Number Queued. The total number of messages
which have been placed on the queue.
Server Requests/second. How many server requests were
handled per second.
Workload/second. The workload is a weighted
measure of the server requests. Some requests could
have a different weight than others. By default, the
workload is always 50 times the number of requests.
Workstation Handler
(WSH)
Bytes Received/sec. The total number of bytes
received by the workstation handler, per second.
Bytes Sent/sec. The total number of bytes sent back
to the clients by the workstation handler, per
second.
Messages Received/sec. The number of messages
received by the workstation handler, per second.
Messages Sent/sec. The number of messages sent
back to the clients by the workstation handler, per
second.
Number of Queue Blocks/sec. The number of times
the queue for the workstation handler blocked, per
second. This gives an idea of how often the
workstation handler was overloaded.
Monitor Measurements
Part XIII
Application Component Monitoring
392
393
46
Introducing Application Component
Monitoring
Using Performance Centers Application Component monitors, you can
monitor the Microsoft COM+ server.
In order to obtain this data, you need to activate the online monitor for the
server and specify which resources you want to measure before executing
the load test.
The procedures for selecting monitor measurements and configuring the
monitors vary according to server type.
Chapter 46 Introducing Application Component Monitoring
394
395
47
Microsoft COM+ Monitoring
The Microsoft COM+ monitor provides information about the resource
usage of the Microsoft COM+ server during load test runs.
Configuring the Microsoft COM+ Server Monitor on page 396
Microsoft COM+ Performance Counters on page 398
To monitor the Microsoft COM+ server performance, you must first set up
the Microsoft COM+ server on the server machine.
Before configuring the monitor, on the Microsoft COM+ server machine,
run the <ComPlus_server_side.exe> file to configure the server for
monitoring.
Before running the Microsoft COM+ monitor over a firewall:
Make sure that the MI Listener is installed on any machine outside of the
firewall. For installation instructions, see the HP Performance Center
System Configuration and Installation Guide.
Configure the Performance Center agent on the server machine. For
more information, see the HP Performance Center System Configuration and
Installation Guide.
Configure the monitor as described in Configuring the Microsoft COM+
Server Monitor below.
Chapter 47 Microsoft COM+ Monitoring
396
Configuring the Microsoft COM+ Server Monitor
After you have set up the Microsoft COM+ server on the server machine,
you can specify which measurements and resources you want the Microsoft
COM+ monitor to measure. You select these counters using the Performance
Center Hosts monitor configuration page.
To configure the COM+ monitor:
Profile.
Click Monitors.
2 Under Application Component Graphs, select Microsoft COM + and click
3 Type the monitor information:
397
For regular monitoring without a firewall, type the name or IP address of
the server whose resources you want to monitor.
To connect to the monitor over a firewall, type the server name or IP
address of the machine you want to monitor, according to the following
format:
where server machine key is the unique key that you chose when
configuring the firewall Agent on the server machine.
For example:
5 Select the objects, counters, and instances that you want to monitor. The
instances are relevant only if multiple instances of the highlighted counter
are running.
<MI Listener machine>:<server machine key>
12.12.12.3:serverid
398
For a list of available performance counters, see Microsoft COM+
profile.
Microsoft COM+ Performance Counters
The following tables describe the default counters that can be measured:
Authentication Metrics
Application Events
Authenticate Frequency of successful method call level
authentication. When you set an authentication
level for an application, you determine what degree
of authentication is performed when clients call
into the application.
Authenticate Failed Frequency of failed method call level
authentication.
Activation Frequency of application activation or startup.
Shutdown Frequency of application shutdown or termination.

399
Thread Events
Transaction Events
Thread Start Rate at which single-threaded apartment (STA)
thread for application have been started.
Thread Terminate Rate at which single-threaded apartment (STA)
thread for application have been terminated.
Work Enque Event sent if a work is queued in single thread
apartment object (STA). Note: These events are not
signaled/sent in Windows Server 2003 and later.
Work Reject Event sent if a work is rejected from single thread
apartment object (STA). Note: These events are not
signaled/sent in Windows Server 2003 and later.
Transaction Duration Duration of COM+ transactions for selected
application.
Transaction Start Rate at which transactions have started.
Transaction Prepared Rate at which transactions have completed the
prepare phase of the two-phase protocol.
Transaction Aborted Rate at which transactions have been aborted.
Transaction Commit Rate at which transactions have completed the
commit protocol.
400
Object Events
Object Life Time Duration of object existence (from instantiation to
destruction).
Object Create Rate at which new instances of this object are
created.
Object Destroy Rate at which instances of the object are destroyed.
Object Activate Rate of retrieving instances of a new JIT-activated
object.
Object Deactivation Rate of freeing JIT-activated object via SetComplete
or SetAbort.
Disable Commit Rate of client calls to DisableCommit on a context.
DisableCommit declares that the objects
transactional updates are inconsistent and cant be
committed in their present state.
Enable Commit Rate of client calls to EnableCommit on a context.
EnableCommit declares that the current objects
work is not necessarily finished, but that its
transactional updates are consistent and could be
committed in their present form.
Set Complete Rate of client calls to SetComplete on a context.
SetComplete declares that the transaction in which
the object is executing can be committed, and that
the object should be deactivated on returning from
the currently executing method call.
Set Abort Rate of client calls to SetAbort on a context.
SetAbort declares that the transaction in which the
object is executing must be aborted, and that the
object should be deactivated on returning from the
currently executing method call.
401
Method Events
Method Duration Average duration of method.
Method Frequency Frequency of method invocation.
Method Failed Frequency of failed methods (i.e. methods that
return error HRESULT codes).
Method Exceptions Frequency of exceptions thrown by selected
method.
402
Part XIV
Middleware Performance Monitoring
404
405
48
Middleware Performance Monitoring
Using Performance Center's Middleware Performance Tuxedo monitor, you
can monitor Tuxedo servers during a load test run and isolate server
performance bottlenecks. You can also monitor the performance of IBM
WebSphere MQ Series servers using the SiteScope monitor.
Configuring the Tuxedo Monitor on page 407
Tuxedo Monitor Measurements on page 410
Note: For information about monitoring the IBM WebSphere MQ Series
servers using SiteScope, see Configuring the SiteScope Monitor on
page 92.
Chapter 48 Middleware Performance Monitoring
406
The Tuxedo monitor allows you to measure and view the performance of
your Tuxedo application server.
If Tuxedo 7.1 or later is installed, you can monitor more than one Tuxedo
application server at a time. However, if Tuxedo 6.5 or earlier is installed on
the machine, you can monitor only one Tuxedo application server at a time.
Note: The Tuxedo workstation client should be installed on local drive. You
cannot install the client on the mapped network drive.
Before you configure the monitor, perform the following:
Check that a Tuxedo workstation client (not a native client) is installed
on the Performance Center Host and on the Utility Server. Use a Tuxedo
6.x client if a Tuxedo 6.x server is used, and Tuxedo 7.1 or later client if a
Tuxedo 7.1 or later server is used. If you use a Tuxedo 6.5 or earlier server,
you can still use a Tuxedo 7.1 or later client to monitor it, provided that
you set the WSINTOPPRE71 environment variable to yes.
Note: A Tuxedo workstation client communicates with the application
server over the network, and is not required to run the Tuxedo
application server on the same machine. A native client can only
communicate with the Tuxedo application server if it is part of the
relevant Tuxedo domain.
Define the Tuxedo environment variables on both the Performance
Center Host and Utility Serverset the TUXDIR variable to the Tuxedo
installation directory, (for example, V:\environ\32\Tuxedo8.0), and add the
Tuxedo bin directory to the PATH variable.
407
Check that the workstation listener (WSL) process is running. This
enables the application server to accept requests from workstation
clients. The address and port number used to connect to the application
server must match those dedicated to the WSL process.
Note: For information on configuring the WSL, see the BEA Tuxedo Web
site (http://edocs.beasys.com/tuxedo/tux81/rf5/rf5101.htm#1534543).
Configuring the Tuxedo Monitor
After you install the client add-in on the Performance Center Host, you
specify which measurements and resources you want the Tuxedo monitor to
measure.
To configure the Tuxedo monitor:
Profile.
Click Monitors.
408
2 Under Middleware Performance Graphs, select TUXEDO and click Next. The
3 Type the details of the server whose resources you want to monitor:
a Type the name or IP address of the server whose resources you want to
monitor.
Note: If you are using multiple instances of the Tuxedo monitor on the
same machine, then type the port number of each Tuxedo monitor in
order to distinguish one instance from another. The entry should have
the following format: <machine name>:<port number>.
b Type the Tuxedo User Name, Password, and Client name.
Logon information is located in the Logon section of the tpinit.ini file in
the recorded scripts directory. Rather than type the values manually, use
the Browse button and select the tpinit.ini file from a recorded script. You
can also determine the client name from the lrt_tpinitialize statement in
the recorded script.

409
In the following example of a tpinit.ini file, the Tuxedo monitor was
configured for a server named psft1 using port 7000, and a client named
bankapp. The logon user name was PS and the password was PS.
profile.
[Logon]
LogonServername=//psft1:7000
LogonUsrName=PS
LogonCltName=bankapp
LogonGrpName=
LogonPasswd=PS
LogonData=
410
Tuxedo Monitor Measurements
The following table describes the default counters that can be measured. Pay
particular attention to the following measurements: %Busy Clients, Active
Clients, Busy Clients, Idle Clients, and all the queue counters for relevant
queues.
Monitor Measurements Measurements
M
a
c
h
i
n
e
% Busy Clients The per cent of active clients currently
logged in to the Tuxedo application server
which are waiting for a response from the
application server.
Active Clients The total number of active clients
currently logged in to the Tuxedo
application server.
Busy Clients The total number of active clients
application server that are waiting for a
response from the application server.
Current Accessers Number of clients and servers currently
accessing the application either directly
on this machine or through a workstation
handler on this machine.
Current Transactions Number of in use transaction table entries
on this machine.
Idle Clients The total number of active clients
application server that are not waiting for
a response from the application server.
Workload Completed/
second
The total workload on all the servers for
the machine that was completed, per unit
time.
Workload Initiated/
second
The total workload on all the servers for
the machine that was initiated, per unit
time.
411
Q
u
e
u
e
% Busy Servers The per cent of active servers currently
handling Tuxedo requests.
Active Servers The total number of active servers either
handling or waiting to handle Tuxedo
requests.
Busy Servers The total number of active servers
currently busy handling Tuxedo requests.
Idle Servers The total number of active servers
currently waiting to handle Tuxedo
requests.
Number Queued The total number of messages which have
been placed on the queue.
S
e
r
v
e
r
Requests/second How many server requests were handled
per second.
Workload/second The workload is a weighted measure of the
server requests. Some requests could have
a different weight than others. By default,
the workload is always 50 times the
number of requests.
W
o
r
k
s
t
a
t
i
o
n

H
a
n
d
l
e
r

(
W
S
H
)
Bytes Received/sec The total number of bytes received by the
workstation handler, per second.
Bytes Sent/sec The total number of bytes sent back to the
clients by the workstation handler, per
second.
Messages Received/sec The number of messages received by the
workstation handler, per second.
Messages Sent/sec The number of messages sent back to the
clients by the workstation handler, per
second.
Number of Queue
Blocks/sec
The number of times the queue for the
workstation handler blocked, per second.
This gives an idea of how often the
workstation handler was overloaded.
Monitor Measurements Measurements
412
Part XV
Troubleshooting Monitors
414
415
49
Troubleshooting Monitors
This chapter provides information on troubleshooting problems related to
Performance Center monitors.
Cannot Create a Monitor on page 415
Cannot Save Measurements on page 417
Port Required for J2EE Monitor Is Already Taken on page 418
J2EE Monitor Initialization Errors on page 419
Cannot Save SiteScope Monitors with Many Counters on page 419
Cannot Create a Monitor
Problem description: You are unable to create a specific monitor.
Troubleshooting
1 Make sure that the user trying to access the machine to monitor exists on
this machine and has proper privileges. Usually IUSR_METRO will be the
user attempting to monitor.
For example, if you add the Windows Resources monitor, the user
IUSR_METRO must exist on the machine to be monitored and must have
proper access rights.
Chapter 49 Troubleshooting Monitors
416
2 Some monitors require certain clients to reside on the Utility Server and the
Controller.
For BroadVision, JDK must be installed.
For WebLogic, JDK and the weblogic.jar file must be present.
For Oracle, the Oracle Client must be installed.
For DB2, the DB2 administrative client must be installed. (OS390 is not
supported.)
3 Make sure the agent required for a particular monitor is running on the
machine that you want to monitor.
4 For COM+, make sure that the COM+ probe is running on the machine that
5 Some monitors require access to a particular URL, from both the
Performance Center utility server and the Controller, for monitoring
purposes.
For Apache, this URL is: http://<server>:<port>/server-status:auto
For Websphere 4.x & 5.x, this URL is: http://<server>:<port>/wasPerftool/
ervlet/perfservlet
417
Cannot Save Measurements
Problem description: During configuration of a monitor, after you select the
measurements that you want to monitor and you click Save, instead of
saving the monitor, Performance Center goes back to the Choose Monitor
page.
Troubleshooting
This problem usually occurs when you select a large number of
measurements. To resolve this issue, you need to increase the value of the
AspMaxRequestEntityAllowed property in the IIS metabase. This property
specifies the maximum number of bytes allowed in the entity body of an
ASP request.
Note: For more information about the AspMaxRequestEntityAllowed
property, see the Microsoft Terminal Services documentation: http://
www.microsoft.com/technet/prodtechnol/WindowsServer2003/Library/IIS/
a6401b5e-c902-4035-90aa-ee46c270d357.mspx?mfr=true, or http://
support.microsoft.com/kb/327659
To increase the value of the AspMaxRequestEntityAllowed property:
1 Run the following command from the command line:
where <IIS_installation_drive> is the hard disk where IIS is installed.
2 To view the current value of the AspMaxRequestEntityAllowed property, run
the following command:
cd <IIS_installation_drive>:\inetpub\adminscripts.
cscript adsutil.vbs get w3svc/ASPMaxRequestEntityAllowed.
418
3 To increase the value of the AspMaxRequestEntityAllowed property, run the
following command:
where <maximum value> is the maximum value, in bytes, that you want to
set for the AspMaxRequestEntityAllowed property.
4 Run the following command:
5 To verify that the value of the AspMaxRequestEntityAllowed property has
changed, run the following command again:
Note: If you are still not able to save the measurements in the monitor
profile, increase the value of the AspMaxRequestEntityAllowed property
further.
For more information about this issue, see the Microsoft support Web site:
http://support.microsoft.com/kb/327659
Port Required for J2EE Monitor Is Already Taken
Problem description: The J2EE monitor communicates with Performance
Center, by default, using port 2004 and this port is already taken.
Troubleshooting
Select another port as follows:
1 On the application server machine, open the <J2EE Monitor installation
directory>\dat\monitor.properties file and change the port number
specified for the webserver.monitor.port property.
cscript adsutil.vbs set w3svc/ASPMaxRequestEntityAllowed <maximum value>.
iisreset.
cscript adsutil.vbs get w3svc/ASPMaxRequestEntityAllowed.
419
2 On the Performance Center host machine, open the <Performance Center
installation directory>\dat\monitors\xmlmonitorshared.ini file and change
the port number specified in mon_j2ee section under the DefaultPort key.
J2EE Monitor Initialization Errors
Problem description: You are receiving application server initialization
errors, such as UnsupportedClassVersionError, NoSuchMethodError, or
NoClassDefFoundError.
Troubleshooting
There might be a conflict between the JDK version specified using the J2EE
Monitor Initializer, and the actual JDK version used in the application server
launch.
Make sure that you selected the correct JDK that is currently being used by
the application server. Note that if you switched the application server to
work with a different JDK, you must run the J2EE Monitor Initializer again.
Cannot Save SiteScope Monitors with Many Counters
Problem description: You get an error when trying to save a SiteScope
monitor profile where the monitor has many counters.
Troubleshooting
This error occurs when there are more than 160 counters on the SiteScope
monitor. The error occurs because the value of the IIS parameter,
ASPMaxRequestEntityAllowed, which defines the maximum size of the Asp
page body section, is not high enough. To solve this error, increase the value
of the parameter.
To increase the value of the IIS parameter:
1 Log in to the User Site server machine.
2 On the Command line, change the directory to the IIS installation directory,
C:\Inetpub>\AdminScripts.
420
3 Run the following commands from the user server and from the utility
a Check the current parameter value.
b Set a higher value for the parameter:
c Check that the parameter value was updated.
cscript adsutil.vbs get w3svc/ASPMaxRequestEntityAllowed
cscript adsutil.vbs set w3svc/ASPMaxRequestEntityAllowed 8000000
cscript adsutil.vbs get w3svc/ASPMaxRequestEntityAllowed
421
A
Adobe Reader 16
Antara FlameThrower monitor 61
performance counters 63
Apache monitor 135
Application Component monitors 391, 393
Microsoft COM+ monitor 396
Application Deployment Solutions monitors
345
Citrix MetaFrame XP 347
Ariba monitor 161
ASP monitor 209
ATG Dynamo monitor 167
performance monitors 170
Auto Discovery mode 328
B
BEA WebLogic monitor 231
BroadVision monitor 174
servers and services 177
C
Check Point FireWall-1 monitor 125
Citrix MetaFrame XP monitor 347
ColdFusion monitor 188
Connections graph 55
Connections per Second graph 55
custom queries, Oracle monitor 307
D
Data Points graph (online) 41
Database Server Resource monitors 271, 297
DB2 monitor 281
SQL Server monitor 273
Sybase monitor 311
DB2 monitor 281
documentation updates 19
documentation, online 16
E
ERP/CRM Server Resource monitors 359
PeopleSoft (Tuxedo) monitor 386
SAP Portal monitor 367
SAPGUI monitor 362
Siebel monitor 374
Siebel Server Manager monitor 381
Error Statistics graph 40
F
Firewall Server monitors 125
firewalls
network monitoring 121
H
Hits per Second graph 50
HP Software Support Web site 18
HP Software Web site 18
HTTP Response per Second graph 51
Index
Index
422
I
IIS monitor 139
iPlanet (NAS) monitor 196
iPlanet (SNMP) monitor 149
iPlanet/Netscape monitor 144
J
J2EE
EJB information retrieval 328
examples of modifying application
server configurations 332
JDBC information retrieval 327
J2EE monitor 323, 329
activating on client machine 328
initial configuration settings 326
installing on application server 324
Java Performance monitors
J2EE monitor 323
K
Knowledge Base 18
M
Microsoft ASP monitor 209
Microsoft COM+ monitor 396
Microsoft IIS monitor 139
Middleware Performance monitors
Tuxedo monitor 407
monitoring using SiteScope 59
monitors
Application Component 391, 393
Application Deployment Solutions
345
Database Server Resources 271, 297
ERP/CRM Server Resources 359
Firewall Server 125
Java Performance 323
Run-Time 39
System Resources 59, 61, 75, 83, 91,
97, 103
Transaction 43
Web Application Server Resources
159, 161, 209
Web Resources 49
Web server resources 131
N
Network Delay monitor 113
Network monitor 111
configuring 117
determining bottlenecks 111
monitoring over a firewall 121
on UNIX 113
O
online documentation 16
online graphs
Connections 55
Connections per Second 55
Error Statistics 40
Hits per Second 50
HTTP Response per Second 51
Pages Downloaded per Second 53
Retries per Second 54
Throughput 50
Total Transactions per Second
(Passed) 44
Transaction 43
User-Defined Data Points 41
online monitors
Network 111
Network Delay 113
online resources 18
online transaction monitoring
adding transactions 45
graphs 43
setup 44
online Web server resource monitoring,
using a proxy server 132
Index
423
Oracle Database Resource monitor 302
custom queries 307
Oracle9iAS 219
Oracle9iAS HTTP monitor 213
P
packets 111
Pages Downloaded per Second graph 53
PeopleSoft (Tuxedo) monitor 386
proxy server 132
R
Retries per Second graph 54
rsh, connection for UNIX Network monitor
115
rstatd process
activating 104
resource monitors 104
Running Vusers graph
Error state 40
Finished state 40
Ready state 40
Running state 40
Run-Time graphs 39
S
SAP CCMS resources, monitoring using
SiteScope 91
SAP Portal monitor 367
SAPGUI monitor 362
Server Resources monitor 98
Services.UserDataPoint(Value,Name) 41
Siebel monitor 374
Siebel Server Manager monitor 381
SilverStream monitor 223
SiteScope monitor 92
SiteScope Oracle9iAS monitor
SNMP Resources monitors 59, 83
configuring SiteScope SNMP by MIB
monitor 87
monitoring using SiteScope 91
SQL Server monitor 273
SSL Connections per Second graph 56
Sybase monitor 311
System Resource monitors 59, 61, 75, 83, 91,
97, 103
Antara FlameThrower monitor 61
Server Resources monitor 98
SiteScope monitor 92
SNMP Resources monitor 83
UNIX Resources monitor 106
Windows Resources monitor 76
T
Throughput graph 50
Total Transactions per Second (Passed) graph
44
Transaction monitor 43
transactions
Transaction monitor 43
Transaction Response Time graph 43
Transactions per Second (Failed,
Stopped) graph 44
Transactions per Second (Passed)
graph 44
Troubleshooting and Knowledge Base 18
Tuxedo monitor 407
U
UNIX Resources monitor 106
UNIX, activating rstatd 104
updates, documentation 19
User-Defined Data Points graph 41
Index
424
V
Vuser status
Run-Time graphs 40
Vusers with Error Statistics graph 40
W
Web Application Server Resource monitors
159, 161, 209
Ariba monitor 161
ATG Dynamo monitor 167
BroadVision monitor 174
ColdFusion monitor 188
iPlanet (NAS) 196
Microsoft ASP 209
Oracle9iAS HTTP monitor 213
SilverStream monitor 223
WebLogic (JMX) monitor 231
WebSphere (EPM) monitor 267
WebSphere monitor 238
Web Resource monitors 49
Web Server Resource monitors 131
Apache monitor 135
iPlanet (SNMP) monitor 149
iPlanet/Netscape monitor 144
Microsoft IIS monitor 139
WebLogic (JMX) monitor 231
WebSphere monitor 238
EPM 267
Windows Resources monitors 76

PC Monitors

Uploaded by

Copyright:

Available Formats

PC Monitors

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

PC Monitors

Uploaded by

Copyright:

Available Formats

HP Performance Center

for the Windows operating systems

You might also like