Question

A team runs a CodePipeline with Source (CodeCommit), Build (CodeBuild), and Deploy (CodeDeploy to an EC2 staging fleet) stages. They want integration tests to run against the deployed staging environment, and the pipeline must automatically stop and NOT promote to the production deployment stage if any integration test fails. The tests are executed by a separate CodeBuild project that returns a non-zero exit code on failure. Which approach best implements this requirement with the least custom code?

Accepted Answer

Add a test action using the CodeBuild provider in a stage placed after the staging Deploy stage and before the production Deploy stage; a failed build fails the stage and halts pipeline progression.

Answer

Configure a CloudWatch alarm on the CodeBuild test project and use an EventBridge rule to invoke a Lambda function that calls StopPipelineExecution when the alarm fires.

Answer

Add the test commands to the buildspec of the staging Deploy stage's CodeDeploy AfterInstall lifecycle hook so failures block the deployment.

Answer

Enable a manual approval action before the production Deploy stage and have engineers review the test results in the CodeBuild console before approving.