A serverless application uses CloudFront, API Gateway, and Lambda. Deployments are currently done by creating a new Lambda version and running a CLI script to update. If the new version has errors, another CLI script redeploys the previous version. The company wants faster deployments and quicker detection and rollback when errors occur. Which approach accomplishes this?

AWS SAM (Serverless Application Model) simplifies serverless deployments and integrates seamlessly with AWS CodeDeploy for Lambda. This integration allows for canary or linear traffic shifting, gradually moving traffic to the new Lambda version. Pre-traffic and post-traffic hooks enable automated testing before and after traffic shifts. Crucially, CodeDeploy can automatically roll back to the previous stable version if CloudWatch alarms (e.g., error rates) are triggered, ensuring quick detection and recovery from errors. The other options are less effective. Nested CloudFormation stacks with change sets offer rollback but lack automated traffic shifting and testing hooks. Combining CLI scripts improves efficiency but still requires manual error detection and rollback. Deploying a new API Gateway endpoint and changing CloudFront origins is a more complex and disruptive process for Lambda-specific changes, lacking the granular control and automation of CodeDeploy.