What happened?

When a Node becomes NotReady, the Pods on it first transition to NotReady (triggering a reconcile where they aren't yet terminating), and later get deleted by the node controller (receiving a deletionTimestamp).
However, the PCLQ controller does not trigger a new reconcile when the deletionTimestamp is added. This causes the Pods to remain in a terminating state without being replaced, resulting in the actual replica count dropping below the expected count and affecting service availability.

What did you expect to happen?

When Pods on a NotReady node are marked for deletion (receive a deletionTimestamp), the PCLQ controller should detect this change and trigger a reconciliation. It should then promptly create new replacement Pods to maintain the desired replica count and ensure service availability.

Environment

• Kubernetes version
• Grove version
• Scheduler details
• Cloud provider or hardware configuration
• Tools that you are using Grove together with
• Anything else that is relevant