CKA — Troubleshooting
Chapter Overview
Troubleshooting is worth 30% of the CKA exam — the single largest domain. If you master this chapter, you’re nearly a third of the way to passing.
On the real exam, you’ll face clusters with broken Pods, unresponsive nodes, misconfigured control plane components, and networking failures. You need to quickly identify what’s wrong and fix it — often within a few minutes per question.
This chapter teaches you the systematic debugging approach that experienced Kubernetes administrators use every day. You’ll learn to recognize symptoms, know exactly which commands to run, and apply the right fix on the first try.
What You’ll Learn
| # | Topic | What’s Covered |
|---|---|---|
| 1 | Troubleshoot Pods | CrashLoopBackOff, ImagePullBackOff, Pending, OOMKilled |
| 2 | Troubleshoot Nodes | NotReady, disk/memory/PID pressure, kubelet |
| 3 | Troubleshoot Control Plane | API server, scheduler, etcd, certificates |
| 4 | Troubleshoot Networking | DNS, Services, NetworkPolicies, kube-proxy |
| 5 | Key Debugging Commands | kubectl, journalctl, crictl, decision tree |
| 6 | Hands-On Lab | Diagnose and fix broken cluster scenarios |
CKA Exam Relevance
| Domain | Weight |
|---|---|
| Troubleshooting | 30% |
This is the heaviest weighted domain on the entire CKA exam. The exam will present you with clusters that have real problems — broken Pods, unhealthy nodes, failed control plane components, and networking issues. You must diagnose and fix them under time pressure.
Exam Tip: Most troubleshooting questions follow a pattern: observe the symptom → identify the root cause → apply the fix → verify. Practice this loop until it’s second nature.
Prerequisites
Before diving into this chapter, make sure you’re comfortable with:
- Core Concepts — Pods, Deployments, Services, Namespaces
- Cluster Architecture — How the control plane components work together (API server, scheduler, controller manager, etcd)
- Workloads & Scheduling — Resource requests/limits, taints, tolerations, node selectors
- Networking — Services (ClusterIP, NodePort), DNS, NetworkPolicies
- Basic Linux —
systemctl,journalctl, file editing withvi/nano
If any of these feel shaky, revisit the relevant chapters first — troubleshooting builds on concepts from every other domain.
Let’s Get Started
Ready to become a Kubernetes troubleshooting expert? This is the chapter that separates administrators who can follow tutorials from those who can fix real problems under pressure.
