Kiali instability on deployments
The kiali operator pod will frequently delete/clean up kiali instead of reconciling it. The kiali operator pod goes through a handful of ansible tasks in setting up. When these progress (or when they complete) the kiali pod is spawned and all is well. However, I have seen (40%-50% of the time) the kiali operator pod begins with ansible tasks to delete kiali on startup. See the attached logs for success/fail. Cluster type was AKS, but the same thing happens on azure OCP and aws OCP.
Talked to @cmcgrath in mattermost and he believes he may have seen this before.
Edited by Ben Mountjoy