Open mh013370 opened 2 years ago
Evaluate whether it's possible to change the pod restart policy and still have the operator function properly.
This is the ticket, imo. It would also unlock more extensive usage of sidecar containers.
This is affecting arangodb operator as well, would appreciate a fix for it, thanks!
This would require a change in ArgoCD or a custom application health Lua snippet for the NiFi pods: https://argo-cd.readthedocs.io/en/stable/operator-manual/health/
Has anyone got a custom heath lua snippet for this?
Bug Report
ArgoCD will never progress the health of Pod resources beyond Progressing when a Pod's restart policy is either never or onFailure. This is intentional as they assume Pods in this case are always ephemeral.
https://github.com/argoproj/argo-cd/issues/7259 https://github.com/argoproj/argo-cd/issues/7182
There's not currently a workaround and several other operators which behave this way are affected.
Environment
Kubernetes cluster kind: rke2
NiFi version: 1.17.0
Possible Solution
Evaluate whether it's possible to change the pod restart policy and still have the operator function properly.