Kubernetes Troubleshooting

Use this skill to diagnose Kubernetes/Helm failures without automatically turning an application deploy into host repair.

Workflow

1. Classify the request.
   • App workload deploy/status/rollout/chart work: stay on Helm/Kubernetes tools.
   • Host/substrate setup or explicit repair request: use Ansible registry/playbooks.
   • Foundation services for this repo, including control panels, observability sinks, and management services such as Grafana for Kubernetes metrics, are control-node Ansible work unless the user explicitly asks to deploy them inside Kubernetes.
   • Ambiguous stateful service: ask whether lifecycle/storage is host-managed or cluster-managed.