[BUG] k8s-status: Completed Job/CronJob pods flagged as problems (READY 0/1 check)

[ersintarhan]

Plugin

davemhammer/k8s-status

Submission checklist

• I have searched existing issues and confirmed this is not a duplicate.
• I am using the latest available version of Noctalia and of the plugin.
• This is a bug in a community plugin, not in Noctalia itself (those belong in the noctalia repo).

Bug

Completed Job / CronJob pods are shown as problems in the widget, panel and /kube problems launcher.

Root cause

In k8s-status/plugin/service.luau, isProblemPhase() correctly whitelists Running/Succeeded/Completed. However the pod problem check combines it with a readiness ratio:

local problem = isProblemPhase(status) or (want > 0 and have < want) or restartsNum > 5

Completed Job pods report READY 0/1 (their containers have exited), so want > 0 and have < want is true and the pod is flagged as a problem despite STATUS=Completed.

Expected behavior

Finished pods (Completed/Succeeded) should not be counted as problems.

Suggested fix

Skip the readiness-ratio check for finished phases:

local finished = status == "Succeeded" or status == "Completed"
local problem = isProblemPhase(status)
  or (not finished and want > 0 and have < want)
  or restartsNum > 5

Happy to send a PR if you agree with the approach.

Environment

• Plugin version: current main
• Cluster: k3s / Talos-based (happens with any CronJob/Jobs that leave completed pods)