Monitor health of NFS volumes

1. Proposed title of this feature request

Monitor health of NFS volumes.

2. What is the nature and description of the request?

OCP should let user know when a NFS volume mounted to a Pod becomes unresponsive, e.g. due to NFS server being unavailable. User (or monitoring app) could take corrective measures, e.g. fix the server or restart the pod somewhere where the NFS server can be reached.

3. Why does the customer need this? (List the business requirements here)

Detection of broken / stale NFS mounts would help customers to prevent application downtime.

4. List any affected packages or components.

Kubernetes, NFS CSI driver.

Implementation details

Upstream Kubernetes is working on volume health: https://kubernetes.io/docs/concepts/storage/volume-health-monitoring/

At time of writing this RFE, it's alpha (Kubernetes 1.22) and has very limited usability (it sends Kubernetes event when a volume becomes "unhealthy"). NFS CSI driver does not implement Volume Health at all.

Source: https://bugzilla.redhat.com/show_bug.cgi?id=1903065