How to categorize this issue?
/area ops-productivity
/kind bug
/priority normal
/platform all
/exp intermediate
/topology garden seed
What happened:
When the VPA is enabled for shoots (.spec.kubernetes.verticalPodAutoscaler.enabled=true) then its health is not visible in the ControlPlaneHealthy condition, i.e., the shoot care controller does not check it. This leads to the fact that a broken VPA is hardly detected.
What you expected to happen:
The ControlPlaneHealthy condition for shoots should also reflect the health of the VPA.
How to reproduce it (as minimally and precisely as possible):
Create a shoot with VPA enabled, break it, and see the ControlPlaneHealthy condition to be unchanged.
kubectl version): allShould we outsource the VPA stuff to a ManagedResource?
We already have code for checking MRs, so we can reuse that :)
Not sure, we don't do this for other control plane components as of now, so this would be inconsistent/confusing. I'd suggest to do it similar like it's done for other components now, and then refactor it if we think it makes sense.
/assign @timuthy
/in-progress