AZ-400 Question 171
Select 2You manage a microservices-based application running on Azure Kubernetes Service (AKS) with strict uptime requirements. You plan to implement a canary deployment strategy to ensure minimal downtime when new versions are released. You also need a reliable rollback method if issues arise after a partial rollout. Which two actions should you take to design a resilient canary deployment approach?
- A
Implement a dedicated canary release stage in your CI/CD pipeline that gradually shifts traffic to the new version, using readiness and liveness probes to monitor health.
- B
Use a single deployment slot with a manual swap process, allowing the new version to overwrite the existing version in production immediately.
- C
Rely only on manual user acceptance testing in a staging environment to validate new builds before deploying to production.
- D
Use monitoring and alerting tools (e.g., Application Insights, Azure Monitor) to automatically trigger rollbacks if key metrics degrade during canary release.
Show answer and explanation
Correct answers: A, D
Explanation
Implementing a canary deployment strategy with robust monitoring and health checks is considered a best practice to achieve resilience in production. Gradually shifting traffic with readiness/liveness probes and automated rollbacks minimizes the risk of impacting the entire user base. Azure Monitor and Application Insights are commonly used services to gather metrics and trigger alerts, aligning with Microsoft� recommended canary, blue-green, or rolling deployment strategies as documented in Azure DevOps best practices.
- A. Correct.
Option 1 is correct. A dedicated canary release stage, combined with readiness and liveness probes, allows you to roll out changes incrementally and verify application health. If an issue is detected, you can stop or roll back the canary release without affecting the entire user base.
- B. Incorrect.
Option 2 is incorrect. Using a single deployment slot with a manual swap results in an immediate overwrite, which does not offer an incremental approach. It increases the risk of downtime or service disruption if the new version is problematic.
- C. Incorrect.
Option 3 is incorrect. While manual testing in staging environments is valuable, relying solely on it does not ensure resiliency during live production updates. Without incremental rollout and automated checks, unexpected issues may surface after full deployment.
- D. Correct.
Option 4 is correct. Monitoring and alerting tools that integrate with your pipeline allow you to track performance, error rates, and user experience in real time. If metrics fall below a defined threshold, a rollback can be automatically triggered for a safer, more resilient deployment process.