Troubleshooting

This topic describes the troubleshooting steps for known issues.

Smart Karpenter on OCI Issues

Nodes are not rotating after the rotateAfter duration.

Preliminary Check

Verify if the disruption budget is not blocking the node rotation.

Solution

Review kubectl get nodepool -n smart-scaler -o yaml disruption settings.

There are orphaned nodes in OCI without NodeClaims.

Solution

Provider GC controller (1.2.0+) automatically cleans these up every 10 minutes.

Manual cleanup

Delete instance from the OCI console if required.

The Kubernetes version mismatched after the control plane upgrade.

Solution

Restart the Karpenter pod to trigger version detection and automatic rotation.

The instance listing is slow.

Solution

Add search-resources IAM permission for improved performance.