All things

Cloud, Data, Machine Learning, and more.

bachir | twitter

GCP DevOps Certification Preparation Guide

10 Sep 2022 by dzlab
Professional DevOps Engineer Certification

I recently passed Google Professional DevOps Engineer Certification and, while preparing for it, I went through a lot of resources. I had to review the documentation of many Google Cloud products, and at no point did I feel that one source covered everything I needed.

This article summarizes the resources I found helpful for passing the exam, plus the topics I wish I had spent more time reading about.

Keep in mind that Google updates its services and certification guides very often, so any source other than the official documentation can become outdated quickly. Before booking the exam, always check the current Professional Cloud DevOps Engineer certification page and the official exam guide.

I started my preparation by reading Google Cloud for DevOps Engineers. It is a good read even if it is not focused only on the exam. It covers general DevOps practices, particularly SRE practices as recommended by Google, and many Google Cloud services that a DevOps engineer is expected to know.

It is a very good starting point if you have little knowledge of Google Cloud services and DevOps. Google also recommends the Site Reliability Engineering book, and you can find more SRE resources from Google on sre.google.

Exam at a glance

The exam is close in difficulty to other Google professional certification exams, but the scope is broad because it touches organization setup, infrastructure, CI/CD, SRE, observability, security, and cost optimization.

The current official guide organizes the exam around five areas:

Preparation strategy

The most useful preparation path for me was:

  1. Read the official exam guide and use it as a checklist.
  2. Review the Google Cloud services listed in each domain.
  3. Build or at least mentally trace an end-to-end deployment path: source repository, build, artifact storage, deployment, monitoring, alerting, rollback, and cost review.
  4. Study SRE concepts separately from product documentation. The exam tests both product knowledge and operational judgment.
  5. Practice scenario questions. Most questions are less about memorizing commands and more about choosing the safest, most maintainable, and most Google-recommended option.

If you only read product pages, the preparation can feel endless. Try to connect each product to an operational decision: when to use it, what problem it solves, how it fails, how to secure it, and how to troubleshoot it.

Organization and environments

You should understand how to bootstrap and maintain a Google Cloud organization for multiple teams and environments.

Resource hierarchy

Know how organizations, folders, projects, and resources fit together. A common pattern is to separate projects by application and environment, for example app-dev, app-staging, and app-prod, then apply policies at the folder or organization level.

Important topics:

For the exam, be comfortable with questions where several answers technically work but only one keeps security, auditability, and future growth under control.

Networking

DevOps questions can involve network architecture, especially when deployments span many projects or connect to existing environments.

You should know:

Infrastructure as code

Infrastructure as code is a major topic. You should know how to automate infrastructure changes, review them, and apply them consistently across environments.

Useful resources:

For the exam, remember that infrastructure changes should be versioned, reviewed, automated, and observable. Avoid answers that rely on manual console changes for repeatable production operations.

Development environments

You may see questions about creating secure and repeatable development environments. Know the purpose of:

The important idea is that developers should get the right tools and access without manually provisioning insecure or inconsistent machines. For AI-assisted development and operations, understand where these tools can help with code, logs, metrics, and troubleshooting, but do not use them as a substitute for knowing the underlying platform.

CI/CD

CI/CD is one of the biggest parts of the exam. You need to understand the full path from source code to a safely deployed production workload.

Cloud Build

Cloud Build is Google Cloud’s managed CI service. You should understand build triggers, build configuration files, private pools, substitutions, service accounts, and logs.

Useful resources:

Good to know: Cloud Build steps share a workspace. This can be useful for passing generated artifacts, test reports, or deployment metadata between steps.

Artifact Registry

Artifact Registry is used to store and manage build artifacts such as container images and language packages.

You should know:

Expect questions that combine Artifact Registry with Cloud Build, Cloud Deploy, vulnerability scanning, Binary Authorization, and IAM.

Cloud Deploy

Cloud Deploy is Google Cloud’s managed continuous delivery service. It helps define delivery pipelines and promote releases through targets such as staging and production.

Read:

The exam can ask you to choose between rebuilding an artifact per environment and promoting the same artifact through environments. In most production pipelines, you want to build once, store the artifact, then promote that artifact through controlled stages.

Deployment strategies

Know the trade-offs between common deployment strategies:

Strategy When it is useful Main trade-off
Rolling deployment Gradually replace old instances with new ones Simple, but rollback can be slower
Blue/green deployment Keep old and new versions separate Safer rollback, but needs extra capacity
Canary deployment Send a small amount of traffic to the new version first Requires good metrics and traffic control
Traffic splitting Shift percentages of traffic between versions Useful for gradual rollout and A/B testing
Feature flags Decouple deploy from release Requires application-level flag management

Useful resources:

Secrets and configuration

Do not put secrets in source code, build logs, or container images. Know how to separate build-time and runtime configuration.

Important services and topics:

For exam answers, prefer short-lived credentials, workload identity, secret managers, and least-privilege service accounts over long-lived keys.

Securing the supply chain

You should understand the security controls around CI/CD:

The general pattern is to build from trusted source, produce signed or verifiable artifacts, scan them, store them in Artifact Registry, enforce policy before deployment, and keep audit logs.

Machine learning pipelines

The current exam guide also mentions CI/CD for machine learning workloads. You do not need to become a machine learning engineer for this exam, but you should understand how ML delivery differs from application delivery: model artifacts, data validation, training pipelines, approval gates, evaluation metrics, and rollback plans matter.

Useful resources:

Compute and runtime platforms

The DevOps exam is not just about CI/CD tools. You also need to know how applications run on Google Cloud and how operational decisions differ by platform.

Cloud Run

Cloud Run is a serverless container platform. It is a good fit when you want to run containers without managing clusters.

Google Kubernetes Engine

GKE is a managed Kubernetes service. It appears often in DevOps scenarios because many CI/CD, security, autoscaling, and observability questions involve Kubernetes.

Compute Engine

Compute Engine still matters for workloads that need VM-level control.

Know when a managed instance group, Cloud Run service, or GKE workload is the better operational fit.

Site reliability engineering

The DevOps certification has a strong SRE flavor. You should be comfortable with reliability concepts and how they influence engineering decisions.

SLIs, SLOs, SLAs, and error budgets

These are core concepts:

Read:

Exam questions often ask what to do when reliability is below target. The safest answer usually reduces risk: pause risky releases, roll back, add capacity, reduce blast radius, or improve observability before continuing.

Capacity and lifecycle management

Know how to plan for quotas, limits, reservations, autoscaling, upgrades, and retirement.

Useful resources:

Incident response

You should know the operational response options when users are affected:

Read:

Observability and troubleshooting

Observability is another large part of the exam. You should understand logs, metrics, traces, dashboards, alerts, and how to use them together.

Logs

Know how Cloud Logging collects, stores, filters, routes, excludes, and exports logs.

Important topics:

You should also know when to route logs to BigQuery, Pub/Sub, or Cloud Storage for analysis, downstream processing, or long-term retention.

Metrics, dashboards, and alerts

Cloud Monitoring is central to SRE and troubleshooting questions.

Make sure you understand alert quality. A good alert should be actionable, tied to user impact or an SLO, and routed to the right team.

Traces and telemetry

Know when tracing helps and how it complements logs and metrics.

Distributed tracing is especially useful when latency or errors involve many services.

Ops Agent and hybrid workloads

Know the role of the Ops Agent for Compute Engine and hybrid workloads.

Troubleshooting approach

For scenario questions, work from symptom to scope:

  1. Is the issue global or limited to a region, version, user segment, or dependency?
  2. Did it start after a deployment, config change, quota change, traffic spike, or dependency failure?
  3. What do logs say?
  4. What do metrics say?
  5. What do traces say?
  6. Can you mitigate user impact before finding root cause?

The exam often rewards mitigation first, then root-cause analysis.

Performance and cost optimization

The last domain combines performance engineering and FinOps. You should know how to collect performance data, use recommenders, and choose cost-effective infrastructure.

Performance

Useful topics:

Questions may ask how to diagnose latency, right-size workloads, or identify bottlenecks. Prefer answers that use measured data rather than guessing.

Cost

Know the common cost levers:

Do not forget observability costs. Logs, metrics, and traces are valuable, but high-cardinality metrics, noisy logs, and unnecessary retention can become expensive. Know how to use filters, exclusions, sampling, and routing.

Topics worth extra review

These are topics I would spend extra time on:

Practice resources

Useful preparation resources:

Certification SWAG

After passing the exam, you can choose one of the official certification swags:

That’s all folks

Check the following preparation tips for passing other Google certifications:

Feel free to leave a comment or reach out on twitter @bachiirc

Related Posts