The Certified Kubernetes Administrator (CKA) exam is a practical test, not a memorization exercise: the Linux Foundation requires a score of 66% or higher to pass, every task is solved from a live command line, and the five curriculum domains are weighted so that troubleshooting alone decides nearly a third of the result. Sound CKA exam preparation follows directly from that structure: spend most of your time breaking and fixing real clusters, learn the documentation you are allowed to open during the test, and rehearse under a clock. This guide covers the current format, the published domain weights, the allowed-resource rules and a weekly routine that converts those facts into a pass.
What the Exam Looks Like
The CKA is an online, proctored, performance-based exam delivered on the PSI Bridge platform with a secure browser, webcam and screen sharing. Candidates have 2 hours to complete the tasks, each of which is a problem on a live cluster rather than a multiple-choice question. The environment tracks the latest Kubernetes minor release — v1.35 on the current exam page — and the Linux Foundation realigns it within weeks of each new version, so confirm the listed version when you book. Enrollment costs $445 and includes one free retake, you receive 12 months of eligibility to schedule the attempt, and the credential stays valid for two years. Your purchase also unlocks two simulations on Killer.sh, each a graded 17-question session with 36 hours of access per activation, which is the closest public rehearsal of exam day.
Domain Weights and Study Time
The CNCF publishes the curriculum and splits it into five weighted domains. Troubleshooting carries the heaviest weight at 30%, with Cluster Architecture, Installation and Configuration next at 25%, followed by Services & Networking at 20%, Workloads & Scheduling at 15% and Storage at 10%. The two heaviest domains together make up more than half of the exam, so your calendar should mirror that: give debugging drills the largest single block, cluster build-out with kubeadm the second largest, and treat Storage as reinforcement rather than a priority.
| Domain | Weight | What to drill |
|---|---|---|
| Troubleshooting | 30% | Failed nodes, kubelet errors, control plane components, pod and service diagnosis |
| Cluster Architecture, Installation & Configuration | 25% | kubeadm cluster creation, RBAC, HA control plane, Helm and Kustomize |
| Services & Networking | 20% | Network policies, ClusterIP, NodePort, LoadBalancer, Ingress, CoreDNS |
| Workloads & Scheduling | 15% | Rolling updates and rollbacks, ConfigMaps and Secrets, autoscaling, affinity |
| Storage | 10% | Storage classes, persistent volumes and claims, access modes, reclaim policies |
We break these weights down further, including how scoring works against the pass line, in our guide to CKA domain weights and the passing score. The short version: if you can diagnose a broken cluster quickly and build one from scratch with kubeadm, you are already above the threshold on the majority of tasks.
Documentation Allowed in the Exam
You do not need to memorize YAML syntax, because the exam permits a short allowlist of documentation opened from inside the exam terminal: the Kubernetes documentation, the Kubernetes blog, the Helm documentation and the Gateway API documentation, plus task-specific links in the quick reference box. Equally important is what is forbidden: external search results, notes, course manuals and any site outside the allowlist. Practice under that constraint from day one. Bookmark the Kubernetes docs, the kubectl cheat sheet and the Gateway API reference in your browser, and drill locating examples quickly, because fast navigation of kubernetes.io is a genuine scoring skill. Write nothing on paper — the desk must be clear — so bookmarks and shell aliases inside the allowed environment are your only accelerators.
Your Week-by-Week Prep Routine
Seven to eight weeks is a realistic runway for a working engineer with container experience. Follow this order:
- Weeks 1–2 — Core objects: pods, deployments, services, ConfigMaps and Secrets on a local cluster created with kubeadm or a managed sandbox.
- Weeks 3–4 — Build from zero: bootstrap clusters with kubeadm, join worker nodes, configure RBAC, install a CNI plugin, and repeat until it is boring.
- Weeks 5–6 — Troubleshooting under time pressure: kill components, corrupt manifests, misconfigure networking, then fix each failure against a stopwatch.
- Week 7 — Full simulations: use both Killer.sh sessions, review every wrong answer, and rebuild your weak spots.
- Exam week — Light drills, a bookmark audit, the PSI system check, and no new topics.
Speed matters as much as accuracy, and shaving seconds per task compounds over the whole exam. Our companion plan on timed drills that build speed for the CKA splits each block into daily sessions with concrete targets. Practice vim and kubectl shortcuts such as the dry-run pipe into a file, and set kubectl aliases so long commands stop eating your clock.
Mistakes That Cost Points
- Studying for multiple choice. The CKA never asks you to identify a correct answer; it asks you to produce a working cluster state.
- Ignoring context. Tasks name a namespace explicitly; applying manifests in the wrong namespace silently fails the check.
- Skipping verification. Graders check state, not intent: end every task with a quick get or describe that proves the change took effect.
- Wasting the simulator. Two graded attempts come with your enrollment fee — skipping both before exam day is throwing away paid preparation.
Sources
- CNCF — Certified Kubernetes Administrator (CKA)
- Linux Foundation T&C Docs — CKA, CKAD and CKS FAQ
- Linux Foundation T&C Docs — Resources Allowed During the Exam