Skip to main content

Troubleshooting

Agent pod not starting​

Check events and logs:

kubectl describe pod -n saviour -l app.kubernetes.io/name=centurion
kubectl logs -n saviour -l app.kubernetes.io/name=centurion

Common causes:

  • RBAC not created — check centurion.rbac.create=true (default)
  • Secret missing — verify global.existingSecret points to an existing Secret
  • Wrong namespace — ensure Secret is in the same namespace as the chart

Cluster not appearing in the UI​

  1. Verify the ingestion key: 
    kubectl get secret saviour-secret -n saviour -o jsonpath='{.data.api-key}' | base64 -d
  2. Check Centurion can reach the backend: 
    kubectl logs -n saviour deploy/saviour-centurion | grep -E "connected|error|failed"
  3. Ensure egress to port 443 is allowed (check firewall rules and NetworkPolicy if enabled).

No metrics from Sentinel​

kubectl logs -n saviour -l app.kubernetes.io/name=sentinel | tail -20

Check that hostPID: true is set (automatic via chart) and the node allows it.

eBPF not working​

  • Check kernel version: uname -r (needs 4.18+)
  • Confirm privileged mode is allowed on the node
  • Check eBPF logs: 
    kubectl logs -n saviour -l app.kubernetes.io/name=sentinel | grep -i ebpf

How to check agent versions​

kubectl get deploy,ds -n saviour \
  -o jsonpath='{range .items[*]}{.metadata.name}{"\t"}{.spec.template.spec.containers[0].image}{"\n"}{end}'

Helm chart version​

helm list -n saviour