AWS Certified DevOps Engineer – Professional — Question 93

A development team manages website deployments using AWS CodeDeploy blue/green deployments. The application is running on Amazon EC2 instances behind an Application Load Balancer in an Auto Scaling group.

When deploying a new revision, the team notices the deployment eventually fails, but it takes a long time to fail. After further inspection, the team discovers the AllowTraffic lifecycle event ran for an hour and eventually failed without providing any other information. The team wants to ensure failure notices are delivered more quickly while maintaining application availability even upon failure.

Which combination of actions should be taken to meet these requirements? (Choose two.)

Answer options

• A. Change the deployment configuration to CodeDeployDefault.AllAtOnce to speed up the deployment process by deploying to all of the instances at the same time.
• B. Create a CodeDeploy trigger for the deployment failure event and make the deployment fail as soon as a single health check failure is detected.
• C. Reduce the HealthCheckIntervalSeconds and UnhealthyThresholdCount values within the target group health checks to decrease the amount of time it takes for the application to be considered unhealthy.
• D. Use the appspec.yml file to run a script on the AllowTraffic hook to perform lighter health checks on the application instead of making CodeDeploy wait for the target group health checks to pass.
• E. Use the appspec.yml file to run a script on the BeforeAllowTraffic hook to perform health checks on the application and fail the deployment if the health checks performed by the script are not successful.

Correct answer: B, E

Explanation

Option B is correct because creating a trigger for deployment failure allows for immediate notification when health checks fail, enhancing responsiveness. Option E is also correct as it runs health checks before allowing traffic, ensuring that issues are detected early. The other options either compromise deployment speed without providing quicker notifications or do not address the need for timely failure alerts.