24 weeks from beginner to a readiness decision
Starting point: no GCP experience; about six months available. Assume 8–10 hours weekly. Confirm your voucher's exact deadline now; this plan does not establish an expiry date. Keep approximately two weeks of contingency if your six-month window permits it. The sequence is based on learning dependencies, not a guarantee that every beginner will be ready in this time.
Weekly rhythm
- Two 90-minute sessions: learn and explain concepts using the study guide and linked documentation.
- Two 90-minute sessions: build, observe, break and fix a lab.
- One two-hour session: practice questions, error log and retrieval from memory.
- Optional one to two hours: revisit a weak topic or repeat a lab without its instructions.
Use one notebook. For every lab save the diagram, commands/configuration, observed output, failure/fix, cleanup evidence and a short explanation. For every missed/guessed question record a corrected rule and when it would not apply. Review after 2, 7 and 21 days.
The sequence
| Week | Learn | Produce evidence |
|---|---|---|
| 1 | Terminal basics, files, processes, HTTP and JSON | Lab 1; explain a response code and find its structured log |
| 2 | Git branches/review, Python basics, container images | Lab 2; build/run an image and explain digest versus tag |
| 3 | Projects, billing, APIs, regions/zones; IAM principals/roles | Lab 3; draw hierarchy and demonstrate scoped access |
| 4 | VPCs, subnets, routes, firewalls, DNS; managed compute choices | Lab 4 design; choose Shared VPC/peering/PSC with reasons |
| 5 | Kubernetes pods, deployments, services, requests/limits | Lab 5 locally; explain scheduling and readiness |
| 6 | Probe failures, logs/events, basic GKE operations | Break and fix lab 5 without copying the fix; foundation checkpoint |
| 7 | Terraform state/plan/apply and environment isolation | Lab 6; review a plan, explain locking/versioning and cleanup |
| 8 | Org policies, federation, impersonation, PAM, residency | Lab 4 security decisions; domain 1 practice and error review |
| 9 | Cloud Build triggers, test stages and untrusted code | Lab 7; demonstrate a failed check preventing the delivery path |
| 10 | Artifact Registry, immutable artifacts and provenance | Record source → build → image digest; review cleanup retention |
| 11 | Cloud Deploy, canary, verification, flags and rollback | Lab 8; verify a known-good revision and restore it after a failure |
| 12 | Secret/configuration management and supply-chain enforcement | Lab 9; compare scan/provenance/attestation; pipeline checkpoint |
| 13 | Logs, metrics, distributions, Ops Agent and collection paths | Lab 10 collection/query tasks; explain cardinality and PII handling |
| 14 | Metrics scopes, PromQL, dashboards, notification routes | Build useful queries; trigger and receive a controlled notification |
| 15 | Trace propagation, sampling and profiles | Lab 10 trace exercise; find latency's critical path |
| 16 | GKE, Cloud Run, database and pipeline troubleshooting | Diagnose four supplied failures; observability checkpoint |
| 17 | SLI/SLO/SLA, budgets and burn-rate calculations | Lab 11; calculate all examples without notes and show units |
| 18 | Capacity, autoscaling, quotas, reservations and DWS | Load-test a small service; explain the three capacity/pricing concepts |
| 19 | Incident mitigation, communication and postmortems | Lab 12 game day; publish a timeline and test one corrective action |
| 20 | FinOps, commitments, Spot, rightsizing, network/log costs | Lab 12 cost review; defend a conservative optimization decision |
| 21 | Combine the five domains; fleets, Gemini and ML pipeline gaps | Capstone and objective checklist; recheck official exam guide |
| 22 | Mock A: first exposure, 120 minutes, no notes | Save first result; review every incorrect and guessed answer |
| 23 | Targeted remediation, then Mock B under the same conditions | Save first result; demonstrate repairs with labs/teach-back |
| 24 | Readiness decision, official samples, exam logistics | Complete the readiness checklist and decide using evidence |
Use practice by domain from week 8 onward; before then, a few relevant questions are fine after learning the concept. Mock A and B are reserved: do not browse their source JSON or answer-review pages beforehand.
Checkpoints
End of week 6: foundations
You can deploy the local service, read its logs, identify its port, describe a pod, and distinguish IAM authorization from network reachability. Explain project/billing account and region/zone. If not, spend the next sessions repairing those foundations rather than accumulating more memorized questions.
End of week 12: delivery
You can trace a change from Git through tests and a digest to deployment, explain which identity acts at each step, keep a secret out of logs, and execute a compatible rollback. You can explain why production migrations complicate rollback.
End of week 16: diagnosis
Given an unavailable app, you can use logs, status/events and metrics to distinguish bad credentials, bad configuration, missing capacity and application failure. You can explain a trace waterfall and why missing metrics are not zeros.
End of week 20: operations
You can define a user-facing SLI, calculate a request budget/burn rate, choose a mitigation and justify a cost optimization without violating reliability. You have practiced at least one failure and recovery yourself.
Capstone for week 21
Use the included service in a small non-production project or a supported training sandbox. Keep a reviewed Git history, a tested build, a registry digest and deployment evidence. Add a health check, structured logs, a useful latency/error signal, a dashboard and tested notification. Document the runtime identity and secret access. Introduce a controlled bad revision or configuration, detect it, mitigate and write a postmortem. Estimate costs and clean up resources.
You should be able to rebuild the essential path using documentation without copying a finished solution. Local Kubernetes work is valuable, but complete real GCP tasks as well before relying on this capstone as certification evidence.
Official learning route
The certification page links the current learning path, official samples and exam guide. Use those as the reference path. Training access/credits can vary; do not assume every lab is free. Learn Associate Cloud Engineer-level fundamentals first, but a separate Associate exam is not a prerequisite purchase.
Your exam may fall in 2027. In weeks 21 and 24, compare the then-current exam guide with coverage-map.md, check current product docs, and repair new gaps. Do not use a date printed on this pack as proof of permanent currency.