Make livenessProbe actually probe for a working runner
What does this MR do?
Makes the livenessProbe for a helm installed gitlab-runner instance actually perform a liveness check.
Why was this MR needed?
Because the previous configuration did not actually check for pod liveness, so the runner could be in a faulty state but the pod never restarts
What's the best way to test this MR?
Install a new runner to the gitlab instance
Remove or disable the runner inside the gitlab GUI
Review the logs & status of the gitlab-runner pod until it has restarted and a new registration was made
What are the relevant issue numbers?
Edited by fiskhest