AZ-400 Question 169
Select 2You are deploying a mission-critical web service to Azure that must remain accessible to users at all times. You want to mitigate the risk of downtime or service disruption when releasing new features. Your deployment plan involves verifying application health after each new version is rolled out and automatically reverting to the previous version if error rates exceed a predefined threshold. Which two strategies should you implement to fulfill these requirements?
- A
Use a blue-green deployment that switches traffic only after the new environment passes health checks
- B
Execute a rolling deployment without any configured rollback mechanism
- C
Implement canary releases with automated monitoring and trigger rollback on elevated errors
- D
Perform a single-stage deployment in production and manually monitor logs post-release
Show answer and explanation
Correct answers: A, C
Explanation
When designing a resiliency strategy for deployment, you want to ensure minimal user impact during new releases and a rapid recovery path in case of issues. Strategies like blue-green and canary deployments use separate or partially isolated environments, perform thorough health checks, and can automatically roll back if predefined error thresholds are met. Microsoft documentation (e.g., 'Azure deployment strategies' at https://learn.microsoft.com/azure/devops/pipelines/release/deployment-strategies) emphasizes the importance of automated rollback to preserve reliability in mission-critical applications.
- A. Correct.
Option 1 is correct. Blue-green deployment involves maintaining two production environments (blue and green). You deploy to the inactive environment, run tests, and flip traffic only if the environment is healthy. In the event of high error rates, you can reroute traffic back to the original environment, making this strategy resilient.
- B. Incorrect.
Option 2 is incorrect. While rolling deployments can reduce downtime, not having a rollback mechanism in place means you cannot quickly revert to a stable state if errors arise, undermining resiliency.
- C. Correct.
Option 3 is correct. Canary releases gradually shift a subset of traffic to the new version and monitor performance metrics. If error thresholds are exceeded, an automated rollback triggers, thus limiting impact on users and improving resiliency.
- D. Incorrect.
Option 4 is incorrect. A single-stage deployment in production exposes your entire user base to potential errors immediately. Relying purely on manual monitoring extends the window of risk and does not align with a proactive rollback strategy.