Troubleshooting monitoring

Metrics are not displayed, Grafana is not available.

The cluster is not created or is in the “Error” status #

1. Cluster dashboard → Events tab. Look for events of type Warning or Error.
2. Conditions — conditions (InfrastructureReady, ControlPlaneReady, etc.) are displayed on the dashboard. Click on the condition for details.
3. Troubleshooting: General troubleshooting for cluster creation, checking the intersection of subnets when creating a cluster.

Nodes in NotReady status #

1. Nodes → select a node → see conditions and events.
2. Causes: kubelet is not running, CNI is not ready, insufficient resources, disk problems.
3. Connect via SSH to the node and check: systemctl status kubelet, crictl ps, kubelet logs.

Pods in CrashLoopBackOff #

1. Workloads → Pods → select pod → Events and Logs tab.
2. Causes: application error, incorrect environment variables, insufficient resources, volume problems.
3. Check kubectl describe pod and kubectl logs.

Metrics or logs are not displayed #

1. Services and repositories → check that shturval-metrics-collector (in the cluster) and shturval-monitoring, shturval-dashboards (in the management cluster) are enabled.
2. For logs - shturval-log-collector in the cluster.
3. Wait 2-5 minutes after turning on - the data does not appear instantly.