Google Professional Cloud DevOps Engineer Question 237
Single answerGoogle Cloud PlatformYour team has deployed a new version of a critical application to production on Google Cloud using a rolling update strategy. Shortly after the deployment, users report significant performance degradation. What is the best rollback strategy to quickly restore the previous stable version while minimizing downtime?
- A
Perform a canary deployment of the previous stable version to test its performance before rolling it out to all instances.
- B
Initiate a rollback using the deployment manager to redeploy the previous stable version to all instances immediately.
- C
Scale down the new version to zero instances and scale up the previous stable version simultaneously.
- D
Manually SSH into each instance and replace the new version with the previous stable version.
Show answer and explanation
Correct answer: C
Explanation
In this scenario, the priority is to quickly restore the application to a stable state to minimize user impact. Scaling down the problematic version and scaling up the previous stable version simultaneously allows for a swift rollback with minimal disruption. Other options, like canary deployments or manual instance updates, are either too slow or inappropriate for urgent rollback scenarios.
- A. Incorrect.
Canary deployment is used for testing changes in a controlled manner, not for immediate rollbacks. While testing the previous version might help, it is not the fastest rollback strategy in this scenario where immediate action is required.
- B. Incorrect.
While using the deployment manager might restore the previous version, it could take longer and introduce complexity if the deployment manager is not already set up for rollback workflows.
- C. Correct.
Scaling down the new version to zero instances and scaling up the previous stable version simultaneously is the quickest way to restore the application to a working state, minimizing downtime and user impact.
- D. Incorrect.
Manually replacing the version on each instance is error-prone and time-consuming, making it unsuitable for a critical and time-sensitive rollback scenario.