Details
-
Bug
-
Resolution: Cannot Reproduce
-
Normal
-
None
-
4.12, 4.11, 4.10
-
None
-
Moderate
-
No
-
False
-
Description
Description of problem:
kube-controller-manager does not fire KubeControllerManagerDown alert when pods are crashlooping
Version-Release number of selected component (if applicable):
How reproducible:
This have been seen in OSD and Rosa clusters in AWS.
Steps to Reproduce:
1. Force kube-controller-manager pods into a crashloop state, on AWS this is as simple as deleting the instance profile foer the control-plane nodes and forcing a redeploy and wait a few for a few deploy attempts 2. 3.
Actual results:
kube-controller-manager operator enters a degraded state
Expected results:
kube-controller-manager operator enters a down state
Additional info: