Question

A team runs a web application on an Amazon EC2 Auto Scaling group behind an Application Load Balancer. They want to use AWS CodeDeploy to deploy new versions to a brand-new set of instances, run smoke tests against them, and only then shift live traffic. If validation fails, the old instances must still be available for an instant rollback with zero downtime. Which CodeDeploy deployment approach meets these requirements?

Accepted Answer

Use a blue/green deployment where CodeDeploy provisions a replacement Auto Scaling group, registers it with the load balancer after validation hooks pass, and keeps the original instances for rollback.

Answer

Use an in-place deployment with the CodeDeployDefault.AllAtOnce configuration so all instances are updated simultaneously.

Answer

Use an in-place deployment with a OneAtATime configuration and rely on Auto Scaling to launch new instances if a health check fails.

Answer

Use a blue/green deployment but terminate the original instances immediately after the new environment is provisioned to reduce cost.