Smlr is a plug for phoenix to compress output to the client, and optionally cache it.

It currently supports Gzip, deflate, br, and Zstd algorithms. The backend is pluggable making it easy for you to add your own compressors.

Local caching of the compressed output is available via cachex, you are able to limit the number of cached items, and or have a ttl for them.

Metrics are implemented via telemetry.

## in your router simply add
plug(Smlr)
## At the beginning of your plug chain, because Plug.Conn.register_before_send is last in first out ordered.

## To compress your websocket traffic add
compress: true

## to your endpoint as described here https://hexdocs.pm/phoenix/Phoenix.Endpoint.html#socket/3-examples

## To turn on the caching support add a  line like this to your config.exs
config :smlr, cache_opts: %{enable: true, timeout: :infinity, limit: nil}

Add the following to the deps block in mix.exs:

{:smlr, "~>1.0.0"}

Configuration can be specified in the opts argument to all Smlr functions, by setting config values with e.g., Application.put_env, or by a combination of the two.

The following configuration options are supported:

• :enable enable or disable the plug entirely. Default true.
• :compressors A list of Modules that follow the Smlr.Compressor behavior. Default [ Smlr.Compressor.Gzip, Smlr.Compressor.Deflate, Smlr.Compressor.Brotli, Smlr.Compressor.Zstd ]
• :compressor_opts A list of tuples, {Module, level} where level is the compression number to set. Default []
• :cache_opts A map of options to be set for cachex. Default cache_opts: %{ enable: false, timeout: :infinity, limit: nil }
• :ignore_client_weight Compress via the order of the compressors list rather than what the client indicated their preference is. Default false
• :all_types Compress all types of data sent into the Plug. Default true
• :types If :all_types is set to false only compress if the content-type matches an item in the list. Default [ "application/atom+xml", "application/javascript", "application/json", "application/xml", "application/xml+rss", "image/svg+xml", "text/css", "text/javascript", "text/plain", "text/xml" ]

Metrics are offered via the Telemetry library. The following metrics are emitted:

• [:smlr, :request, :pass] Client and Server had no compatible compressors, or client did not request compression
• [:smlr, :request, :compress] Compressed response
• [:smlr, :request, :cache] Served compress content from cache
• [:smlr, :request, :cache, :not_started] Cachex was called but had not been started yet.
• [:smlr, :request, :cache, :miss]
• [:smlr, :request, :cache, :hit]
• [:smlr, :request, :cache, :set]

Thanks for considering contributing to this project, and to the free software ecosystem at large!

Interested in contributing a bug report? Terrific! Please open a GitHub issue and include as much detail as you can. If you have a solution, even better -- please open a pull request with a clear description and tests.

Have a feature idea? Excellent! Please open a GitHub issue for discussion.

Want to implement an issue that's been discussed? Fantastic! Please open a GitHub pull request and write a clear description of the patch. We'll merge your PR a lot sooner if it is well-documented and fully tested.

Copyright 2020, Matthew O'Gorman.

This software is released under the MIT License.