k8s-pvc-healthcheck

[stewartshea]

🌟Welcome to the community discussion for the troubleshooting commands defined in: k8s-pvc-healthcheck

🌱This discussion is intended to help grow the community
🛠️Together, let's enhance the usefulness of these troubleshooting commands
🙌 Your participation is crucial! Share your valuable comments and questions here, and don't forget to give a thumbs up if you find the commands helpful and utilize them in your work.

[stewartshea]

Just added a command to check for PVC's only (not PV's) in a Terminating state. It's interesting since a PVC in my test (where I delete the PVC that is still attached to a running pod), won't actually have status.phase set to "Terminating", it's still "Bound" - but the CLI dhows Terminating likely due to the deletionTimestamp annotation which we use in the command.

I wonder if this is specific to this type of scenario and if other scenarios might require additional tuning.

Are there any other scenarios that we might consider testing?

[stewartshea]

Needed:

• raise issue for pods that cannot have utilization checked (e.g. crashlooping)
• SLI to regularly check the condition above as well as the utilization. (every 5m or so?)

@stewartshea (self-assign)

[stewartshea]

This has been added. It could be further enhanced to validate if the pod that has storage should be running. What comes to mind is to determine if the pod belongs to a job or not. For example, a gitlab backup pod raises an issue, but is a job that runs to completion, so it's really an expected behaviour