Log Periodic Client OS-Level Stats

Following on from MB-22696, would it be at all possible to periodically log any useful OS-level stats from the client?

For example, we have just seen a case where a customer's application servers were experiencing timeouts due to increased CPU steal in their client VMs. The onus is generally on Couchbase to prove that Couchbase Server isn't at fault, so any way of gleaning pertinent information from the client could be useful (e.g. CPU load, memory usage, network status).

Appreciate that this would add extra load to the cluster, but this could be optional, and perhaps only stored temporarily and/or aggregated like ETS tables.

Alternatively this could be dumped out periodically to the application log on the client - would at least give us a way of seeing any recent trends on the client server if they do encounter any issues.