Release and Roll Back an AWS Cloud-Native Workload

Table of Contents

1. Challenge

   Prepare the serverless workload and release inputs

   Prepare the healthy release baseline and the inputs for the next deployment:

   • Verify that the prod alias serves version 1 and the workload alarm is OK
   • Inspect the CodeDeploy canary target and request-triggered rollback configuration
2. Challenge

   Release the new version with a CodeDeploy canary

   Start a controlled canary release and confirm that traffic has begun shifting:

   • Create the CodeDeploy deployment from the supplied AppSpec
   • Generate live validation traffic from the workstation
   • Confirm that prod routes approximately 10 percent of requests to version 2
3. Challenge

   Validate release health from deployment and workload evidence

   Evaluate whether the canary meets the release criteria by correlating deployment and workload evidence:

   • Compare CodeDeploy traffic-shift status with the direct version 2 failure
   • Inspect sampled canary failures and the independent CloudWatch alarm
   • Decide whether the release should continue
4. Challenge

   Roll back the failed release

   Stop the unhealthy canary and restore the known-good release:

   • Trigger CodeDeploy's stop-and-rollback action
   • Confirm that the original deployment is Stopped
   • Confirm that the linked rollback deployment is Succeeded
5. Challenge

   Verify recovery and capture the release outcome

   Prove that the workload recovered and preserve the evidence-backed outcome:

   • Confirm that prod serves version 1 with no additional routing weight
   • Verify healthy traffic and the CloudWatch alarm returning to OK
   • Record the linked deployment and recovery evidence in the outcome file