CKAD Study Plan for Kubernetes Developers

Week 3 — design, build and deployment

Goal: the workload-management core, plus Helm and Kustomize.

• Deployments end to end: rolling updates, rollbacks with revision history, scaling, and the difference between recreate and rolling strategies.
• Jobs and CronJobs, including completions, parallelism and backoff limits.
• Labels, selectors and annotations as the connective tissue for everything else.
• Package and customise applications with Helm and Kustomize at the level the curriculum expects — installing, upgrading, overlaying — not chart authorship.

This is also the week to switch your daily work over, where policy allows: use your practice habits on real tickets, because nothing embeds kubectl rollout undo like needing it.

Milestone check: deploy an app, roll out a bad image, diagnose it, roll back, and scale — inside ten minutes.

Week 4 — services, networking and observability

Goal: close the classic developer blind spot.

• Services (ClusterIP, NodePort) and how DNS names resolve between namespaces.
• Ingress rules for host- and path-based routing.
• NetworkPolicies: write one that restricts traffic and then prove it works — the proving is what the exam grades.
• Observability: kubectl logs (including previous-container logs), kubectl describe, events, and debugging Pods that never reach Running.

Milestone check: expose a two-tier app through a Service and Ingress, lock it down with a NetworkPolicy, and debug a deliberately broken deployment a colleague (or you, yesterday) sabotaged.

Week 5 — speed drills and full simulation

Goal: convert competence into pace.

The CKAD is widely regarded as a race; two hours is generous only if you never hesitate. This week, stop learning topics and start drilling them against a stopwatch — set per-task budgets of five to eight minutes and treat overruns as failures even when the YAML was eventually correct. (The drill techniques themselves — imperative fluency, editor setup, skip-and-return triage — are covered in depth in our guide to preparing for the CKAD exam; this plan just tells you when to apply them.)

Take your first full-length timed simulation at the end of the week. Your exam purchase includes two 36-hour sessions of the killer.sh simulator — deliberately harder than most candidates find the live exam — so spend the first session now and save the second for week six. Score it against the 66% pass mark and, more importantly, tag every dropped mark with its domain.

For extra repetitions between simulator sessions, timed CKAD practice questions let you rehearse objective by objective and see which domain keeps leaking marks — analyse the results rather than replaying questions until you remember answers, which teaches you the questions, not Kubernetes.

Milestone check: first full simulation completed, scored, and decomposed into a weak-domain list.

Week 6 — weak domains, second simulation, logistics

Goal: repair, confirm, book.

1. Days 1–3: re-drill only the task types that dropped marks in the simulation. Depth beats breadth now.
2. Day 4: second killer.sh session. You want a score comfortably above 66% on tasks you have not seen before.
3. Day 5: logistics rehearsal. The exam is delivered remotely via the PSI Bridge secure browser — run the compatibility check on your actual machine, clear your desk, remember the single-monitor rule, and re-read the candidate handbook.
4. Day 6: light consolidation only — aliases, doc bookmarks, context-switching habit (kubectl config use-context before every task, without exception).
5. Day 7: rest, then sit the exam.

If the second simulation lands below the pass mark, defer without guilt: your purchase includes a 12-month eligibility window and one free retake, so there is no financial pressure to sit before you are ready. As of 2026 the exam is listed at $445 USD standalone (course and subscription bundles cost more; confirm current pricing on the Linux Foundation's page).

Common ways this plan goes wrong

• Watching instead of typing. Video courses feel like progress; the exam only measures what your fingers know. Cap passive study at one-third of any week.
• Perfect YAML from scratch. If you are still hand-writing full manifests in week four, return to the week-1 imperative drills.
• Ignoring the 25% domain. Developers gravitate to Deployments and Services because they are familiar. Configuration and security is the biggest single slice of the exam.
• Saving all timing pressure for the end. The stopwatch enters in week five here, but per-task time awareness should colour every session from week two. The exam's difficulty is pace, as the assessment of how hard the CKAD exam is makes clear.
• Booking on hope. Two above-threshold simulations are your evidence. One lucky run is not.

Your first week starts tonight

The distance between "I use Kubernetes at work" and "I can pass a two-hour live-cluster exam" is not knowledge — it is repetitions under a clock, and six weeks of them is enough for most working developers. Spin up the practice cluster tonight, generate your first Pod manifest imperatively, and let the plan carry you from there. Pass it, and the credential is valid for two years — and if your ambitions later turn towards cluster operations, the administrator track and its own checklist will be waiting.