AutoscalingRunnerSet gets stuck thinking runners exist when they do not #3821
Labels
bug
Something isn't working
gha-runner-scale-set
Related to the gha-runner-scale-set mode
needs triage
Requires review from the maintainers
Checks
Controller Version
0.9.3
Deployment Method
Kustomize
Checks
To Reproduce
Describe the bug
One of our AutoscalingRunnerSets will often think that runners exist when they do not (the rest of our groups have not been affected). At this point, the github action runs will be pending a runner, but no pod will spin up. When it gets stuck in this state, the only way I've been able to correct is deleting the AutoscalingRunnerSet and re-creating it. Deleting the controller & listener did not help.
Here's an example of the kubernetes commands ran to diagnose and fix (argoCD re-creates the deleted resource for me).
Describe the expected behavior
The AutoscalingRunnerSet should create pods when desired, and not think pods exist that do not.
I'm curious if the pod tracking mechanism is a counter that increments/decrements rather than looking at source of truth (how many pods actually exist)
Additional Context
Controller Logs
Logs from when it was fixed by deleting the AutoscalingRunnerSet and re-creating
The text was updated successfully, but these errors were encountered: