Before I cover the Google Cloud services that implement CI/CD on the Professional Cloud Architect exam, I want to lay out what CI/CD is actually for. The exam expects you to recognize when a CI/CD pipeline is the right answer to an architectural problem, which means understanding the problems CI/CD was invented to solve and the principles it operates on.
CI/CD exists because traditional software delivery had a recurring set of failures. There are four worth naming.
The first is integration hell. Development teams worked in isolation for long stretches, then tried to merge their code with the rest of the team at the end. The result was conflicts, complex merges, and sometimes days of work spent just getting different parts of the codebase to function together again.
The second is late bug detection. Without frequent automated testing, critical bugs were discovered just before release. Fixing a bug at that point is expensive because the code has drifted far from where the bug was introduced, and the release schedule is already locked.
The third is slow releases. Manual deployment processes were complex and risky, so teams released infrequently. A rare release is a stressful release. It requires huge coordination, often causes downtime, and bundles a long list of changes into a single high-risk event.
The fourth is lack of feedback. Developers had to wait a long time to see their code in action. That delay slowed down learning, slowed down iteration, and made it harder to respond to what users actually needed.
CI stands for Continuous Integration. It is about ensuring that code changes are integrated into a shared repository frequently and reliably. CI rests on four practices.
Version control tracks changes to code. It makes collaboration possible and makes it easier to find the change that introduced a bug. An automated build process compiles every code change so that the result is always something runnable. Automated testing validates each change immediately, which is what catches bugs early instead of late. Continuous feedback mechanisms surface the results of those builds and tests back to the developer who made the change, so the loop closes quickly.
The short version: CI automates builds and code testing so that teams can integrate changes frequently and with confidence.
CD stands for Continuous Delivery. It extends CI by automating the release and deployment process. CD has its own four practices.
A deployment pipeline standardizes how code flows from development to production. Environment management standardizes the environments code runs in along the way, so that code behaves consistently from one stage to the next. Release automation reduces the manual effort and the risk in pushing new versions. Deployment strategies like blue/green and canary releases roll out changes in a controlled way, which minimizes risk to end users when something goes wrong.
Blue/green keeps two production environments and switches traffic between them. Canary releases route a small slice of traffic to the new version before sending everyone there. Both exist so that a bad deploy can be rolled back without taking the whole user base with it.
The short version: CD builds on CI by letting teams release software to users quickly and safely, turning updates into low-risk events instead of stressful ones.
The Professional Cloud Architect exam tests whether you can pick services and patterns that match the requirements of a scenario. CI/CD scenarios on the exam tend to come in three shapes. A team is shipping infrequently and needs to release more often. A team is finding bugs late and needs faster feedback. A team needs to deploy without downtime and needs a controlled rollout strategy. Each maps directly to one of the four historical problems above and to one of the eight principles across CI and CD.
When you see a question that mentions any of those failure modes, the answer involves a CI/CD pipeline, and the specific Google Cloud services in that pipeline depend on which principle is being violated. Knowing which problem CI/CD solves is what lets you tell the difference between a question about Cloud Build, a question about deployment strategies in GKE, and a question about artifact storage.
My Professional Cloud Architect course covers CI/CD principles alongside the rest of the architecture and compliance material.
