Question

A CloudOps engineer manages a web application deployed via a CloudFormation stack that includes an Auto Scaling group and an Application Load Balancer. During stack updates, the team has occasionally deployed a bad application version that passes CloudFormation resource creation but causes elevated HTTP 5xx errors in production. The engineer wants CloudFormation to automatically roll back a stack update if the application's error rate spikes during and shortly after the update completes. Which approach meets this requirement with the least custom code?

Accepted Answer

Configure rollback triggers on the stack update referencing a CloudWatch alarm that monitors ALB 5xx errors, and set a monitoring time period after resources are provisioned.

Answer

Add a CreationPolicy with a wait condition to each Auto Scaling instance so CloudFormation waits for a success signal before completing the update.

Answer

Enable stack termination protection and create an EventBridge rule that deletes the stack when the 5xx alarm fires.

Answer

Use a stack policy to deny updates to the Auto Scaling group unless the CloudWatch alarm is in the OK state.