A CKA prep plan only works when it mirrors the exam it targets. The Certified Kubernetes Administrator certification is an online, proctored, performance-based exam that gives you 2 hours to solve hands-on tasks from the command line against a cluster running Kubernetes v1.35, and no amount of video watching replaces hours spent typing commands into a live cluster. The seven-week schedule below front-loads the heaviest domains, builds speed through timed drills, and ends with two full simulator rehearsals so the real session feels like a routine you have already run twice.
Plan the budget alongside the calendar. The exam-only purchase costs $445, the THRIVE-ONE subscription bundle runs $625, and the bundle with the Kubernetes Fundamentals (LFS258) course is $645. Every tier carries the same core entitlements: a 12-month window to schedule, a retake, and two Killer.sh simulator sessions. Engineers who already operate clusters daily can take the exam-only tier; buy a bundle only if you need structured coursework to close foundational gaps before the drills begin.
What the CKA Tests
The exam distributes its weight unevenly across five domains, and that unevenness is the single most useful planning fact you have. Troubleshooting is the heaviest-weighted domain at 30%, while Storage is the lightest at 10%, so diagnosis drills deserve roughly three times the calendar space you give to persistent volumes. Treat the weights below as your default time split, then adjust after your first rehearsal exposes where points actually leak.
| Domain | Weight | Priority drill |
|---|---|---|
| Troubleshooting | 30% | Broken nodes, failed pods, component logs, events |
| Cluster Architecture, Installation & Configuration | 25% | kubeadm installs, upgrades, RBAC, etcd backup |
| Services & Networking | 20% | NetworkPolicies, Ingress, CoreDNS, Gateway API |
| Workloads & Scheduling | 15% | Rolling updates, taints, affinity, autoscaling |
| Storage | 10% | Persistent volumes, claims, storage classes, reclaim policies |
Weeks 1-2: Rebuild Core Skills
Spend the first fortnight rebuilding fluency that may have rusted since you last ran a cluster by hand. Work through this sequence on a disposable environment you can break without consequence, and repeat each step until it stops requiring thought:
- Install a fresh cluster with kubeadm, tear it down completely, and repeat the cycle until every step is muscle memory.
- Deploy workloads with Deployments, StatefulSets, DaemonSets, Jobs and CronJobs, then force and verify a rollback of each type.
- Create Roles, ClusterRoles, RoleBindings and ServiceAccounts, and confirm permissions with kubectl auth can-i before moving on.
- Practice reading component logs and cluster events until diagnosing a failed pod takes under three minutes end to end.
Weeks 3-5: Domain-Weighted Drills
With fundamentals restored, weight every study week to the exam’s own distribution rather than to your comfort zone. Roughly a third of drill time goes to diagnosis scenarios, a quarter to cluster architecture and configuration, a fifth to networking, and the remainder to workloads and storage. For a deeper breakdown of how each domain is scored and what the passing bar demands, read our guide to CKA exam domains, weights and the passing score. Structure each drill as a timed box: one task, one clock, one honest grade against the manifest you meant to produce. Ten graded tasks beat fifty passive tutorials, because the exam measures execution under pressure, not recognition.
Weeks 6-7: Simulator Rehearsals
The final fortnight belongs to full-length rehearsal. Enrollment includes two Killer.sh sessions, and each activation opens a 36-hour access window that starts the moment you launch it, so do not burn both attempts in the same weekend. Each Killer.sh CKA simulator session presents 17 questions drawn from a different question set than the other session, which makes the first attempt a diagnostic and the second a verification of the fixes you made afterward. Run the first rehearsal at the end of week six, grade yourself ruthlessly against the provided solutions, spend the following days drilling only what you missed, then run the second rehearsal under strict exam conditions with no pauses. For the mechanics of running a timed session well, see our walkthrough of CKA exam format, rules and timed drills.
Exam-Day Execution Rules
Performance exams are lost on pacing more often than on knowledge. Fix these rules before exam day so they hold automatically under pressure:
- Set the context first. Several tasks target different clusters; verify and set the kubectl context before touching any object, or your changes land on the wrong cluster.
- Bank the cheap points. Flag heavy tasks, solve the quick ones first, and return to the hard ones with a clear picture of remaining time.
- Generate, do not hand-write. Use dry-run output and imperative commands to scaffold YAML, then edit instead of typing full manifests from scratch.
- Verify everything. Grade your own work the way the exam does: the object must exist, sit in the right namespace, and behave as the task demands.
- Use the open docs deliberately. kubernetes.io documentation is reachable during the exam; know the specific pages you will reach for instead of browsing.
Sources
- Certified Kubernetes Administrator (CKA) — The Linux Foundation
- CKA exam simulator FAQ — Killer Shell