arc-runner-troubleshooting

Queued jobs on ARC runners typically stem from three issues: empty runner labels due to releaseName vs runnerScaleSetName mismatch, cold-start delays from minRunners: 0 configuration, or cluster access failures. Diagnose by validating Helm release names, checking runner label registration, and verifying kubeconfig token expiration.

Empty labels occur when the Helm releaseName does not match the runnerScaleSetName in your ARC configuration. Verify both values align in your Helm values file, redeploy the release, and confirm runners register with expected labels using kubectl to inspect runner pool status.

Scaling faults in ARC result from misconfigured minRunners settings, expired kubeconfig tokens blocking cluster communication, DNS resolution failures, or incorrect custom runner images lacking required tools. Parse GitHub Actions runner data and validate cluster connectivity to identify the root cause.

Runner registration issues stem from access failures, kubeconfig expiration, or pool label mismatches. Validate cluster access via kubectl commands, refresh kubeconfig tokens, verify runner pool configuration against Helm release names, and check for 'unknown' OS detection indicating image problems.

Yes. Cold-start delays occur when minRunners is set to 0, forcing the cluster to provision runners on first job arrival. Adjust minRunner values to keep baseline runners warm, verify scaling policy configuration, and run verification checks to confirm runners are available before job submission.

Validate Helm releaseName against runnerScaleSetName, parse runner registration data to detect empty or 'unknown' labels, verify custom runner images contain preinstalled tools, test cluster kubeconfig access, and run DNS resolution checks to confirm connectivity across runner pools.