You could monitor the histograms of each namespace on each node that is configured on the cluster in the log file. The log output has the following format: histogram dump: {ns-name}-{hist-name} (1344911766 total) msec (00: 1262539302) (01: 0049561831) (02: 0013431778) (03: 0007273116) (04: 0004299011) (05: 0003086466) (06: 0002182478) (07: 0001854797) (08: 0000312272) (09: 0000370715)
The total value is accumulated since the node is started. By tracking change in total count for each histogram (e.g. read, write etc.) - depending on your application, number of cluster nodes, and namespaces configured, you can infer if the cluster is being used or not.