Examkin

Question

Difficulty: MediumRelease Management and Deployment Strategies

A Cloud Architect is establishing a progressive release process using Google Cloud Deploy to update a microservice running on a Google Kubernetes Engine (GKE) cluster. To guarantee reliability and automated health verification before full traffic cutover, place the following deployment pipeline steps in the correct chronological order.

  1. 1Define a Cloud Deploy delivery pipeline specification using Skaffold to declare progressive deployment phases (e.g., 10%, 50%, and 100% traffic allocation).
  2. 2Create a new release in Cloud Deploy to render the Kubernetes manifests and deploy the initial 10% canary workload to the target GKE cluster.
  3. 3Execute automated Cloud Deploy verification jobs against the 10% canary workload to validate latencies and error rates against service level indicators (SLIs).
  4. 4Approve the pipeline promotion gate to advance the release to 100% traffic allocation across all production GKE cluster pods.
  5. 5Retain the previous successful deployment release state in Cloud Deploy to allow immediate single-command rollback if post-promotion anomalies occur.

Answer

The correct sequence starts with pipeline definition, followed by release creation for the initial canary workload, execution of automated verification tests, promotion to 100% traffic, and maintaining the previous release state for rollback readiness.
A progressive delivery strategy requires defining pipeline rules first, deploying to a limited canary tier second, validating health via automated verification third, promoting to full production fourth, and maintaining previous release history for rapid rollback fifth.

Step-by-Step Solution

1
Configure the pipeline declarative manifests.
Cloud Deploy and Skaffold configurations establish the delivery target stages and progressive traffic phases.
Infrastructure and pipeline configuration must exist prior to artifact deployment.
2
Instantiate the release in Cloud Deploy.
Container images and Kubernetes manifests are rendered and deployed to the initial 10% canary phase on GKE.
Creating a release triggers the actual rendering and initial phase deployment.
3
Run automated verification tasks.
Cloud Deploy verification containers execute tests to measure application health and SLIs on the canary workload.
Automated testing ensures defects are caught while blast radius is limited to canary traffic.
4
Promote the release to full traffic cutover.
Traffic shifts completely (100%) to the new workload across the GKE cluster upon successful verification.
Promotion gates validate health before replacing the active production version.
5
Maintain rollback readiness.
The target retains history of the previous stable release manifest.
Preserving previous release states ensures immediate operational recovery if unexpected issues arise post-release.

Key Concept

Progressive Canary Deployment Automation with Cloud Deploy
Rate this question

Topics

Release Management and Deployment Strategies

More questions from this topic