Training material
Debugging Kubernetes hands-on
Finding and fixing failures – from Pending to the control plane, on RKE2 and Talos
An RKE2 cluster with three nodes and a Talos cluster that you break yourself – pending pods, image and start failures, crashes and exit codes, kubectl debug, services, DNS and network policies, Ingress, Gateway and TLS, storage, RBAC and Pod Security, Kyverno and Falco, nodes, the control plane and etcd, webhooks and stuck namespaces, rollouts, autoscaling, jobs and Helm, shell-less nodes on Talos, a competition on a broken cluster and a final exam.
- Kubernetes
- kubectl
- k9s
- RKE2
- Talos
Start the course
Free in your browser, no sign-up
Download for free
PDF and EPUB · edition 2026
Training plan
- Preface
- Day 0: The Training Ground
- Day 1: The Method
- Day 2: k9s
- Day 3: Pods Stuck in Pending
- Day 4: Images and Container Start
- Day 5: Crashes and Restarts
- Day 6: kubectl debug
- Day 7: Services and Endpoints
- Day 8: DNS and NetworkPolicies
- Day 9: Ingress, Gateway and TLS
- Day 10: Storage
- Day 11: RBAC and Pod Security
- Day 12: Faults caused by Kyverno
- Day 13: Falco – alerts and faults
- Day 14: Nodes
- Day 15: Control plane and etcd
- Day 16: Webhooks, APIServices and stuck namespaces
- Day 17: Rollouts, autoscaling, Jobs and Helm
- Day 18: Talos – nodes without a shell
- Day 19: The same faults on Talos
- Day 20: Competition – the broken cluster
- Appendix A: Cheat Sheet
- Appendix B: Final Exam