GCP Professional Cloud Architect Practice Question

An enterprise is containerizing a multi-service retail application and adopting Cloud Build and Cloud Deploy for its CI/CD pipeline. Regulatory policy states that every release must pass black-box system tests in an environment that is configuration-identical to production before any traffic can be directed to production. If the system tests fail, promotion must halt automatically, but the container images should remain available for post-mortem analysis. Which pipeline design best satisfies these requirements while minimizing manual effort?

Configure Cloud Build to build container images and push them to Artifact Registry, then define a Cloud Deploy pipeline with a staging GKE target identical to production. Add a verify job that runs the end-to-end system test suite; Cloud Deploy will promote the release to the production target only if the verify job passes, leaving the images in Artifact Registry even on failure.
Require developers to execute the system test suite on their workstations; once tests pass locally, tag the container images and deploy them to production with manual gcloud commands.
Use Cloud Build to deploy blue/green releases to a single production GKE cluster and rely on Cloud Monitoring SLO-based alerts to detect problems instead of pre-production system testing.
Add a Cloud Build step that promotes the release directly to the production GKE cluster, then runs system tests against production; trigger an automated rollback script if any test fails.

Report Issue

Answer Description

The most efficient approach is to let Cloud Build build the containers and store them in Artifact Registry, then use Cloud Deploy to orchestrate progressive delivery. A staging GKE target that mirrors production provides the required parity. Cloud Deploy's verify (or pre-promotion) job can run the full system test suite; a failure marks the release as failed and prevents automatic promotion to the next (production) target. Because containers are already pushed to Artifact Registry during the build, they remain available for later debugging regardless of test outcome. The other options either test after production deployment, omit an equivalent staging environment, or rely on manual gates, all of which violate the stated policy or add unnecessary human intervention.

Ask Bash

Bash is our AI bot, trained to help you pass your exam. AI Generated Content may display inaccurate information, always double-check anything important.