Hello,
We are using gitlab and gitlab runners on top of GKE.
We noticed that in some cases that though the build container is failing the runner pod does not restart.
Since the restartPolicy of the runner pods is set to “never” (it appears in the runtime env.), and since K8S default is “always” and since we could not find the setting of the restart policy in the helm charts, we have some questions,
- why is the restart policy of the runner pod is set to never? where is it done?
- does gitlab actively restart the pods when identifying issues?
- how does gitlab handles the runner pods during a cluster scale down? is it handled at all?
Thanks so much,
Meir