Maintenance windows

Routine maintenance windows and the operator checks that help distinguish expected change from unexpected drift.

Routine checksChange windows

Before the window

  • Review affected namespaces, workloads and image metadata.
  • Capture current warnings so new warnings stand out clearly afterward.
  • Confirm owner and escalation contacts for the affected areas.

During the window

Use workload and node details to confirm that expected changes are moving in the right direction. Observability should show short-lived warning spikes for normal rollout or restart activity; prolonged warning growth deserves follow-up.

After the window

Check that readiness returned to expected levels, no namespace remains unusually noisy and the registry shows the intended image metadata where applicable.