Currently, we will attempt to return the number of CPUs available to the process, but if sched_getaffinity returns an error, we will default to returning the number of CPUs available to the system host. This is confusing for users of FTDC that are looking at deployments that are running in virtualized environments.

We should look to differentiate the metrics so it's clear whether we are reporting the number of CPUs available to the process or to the host.