Description of problem:
At percent to determine the state of a nova service such as the nova-compute agent
you have to query the service api to determin if its up and then inspect the service logs to ensure there are no errors that would indicate a problem.

This is no trivail to automate or consume in orchstration or monitroing context such as automatically restarting a service if it hangs.

To this end nova should be enhanced to provide meaningful, local, per binary health check that can be consumed by automation,loging and deployment tooling.

https://review.opendev.org/q/topic:per-process-healthchecks

This epic will enable the use of health for service liveness checking in a podifed deployment.