A team has Lambda microservices that read from DynamoDB. Developers manually deploy code after tests. They now require cloud-based automated testing and deployment, with traffic to new versions shifted incrementally after deployment. Which approach delivers these requirements while maximizing developer velocity?

The correct approach leverages CodePipeline for end-to-end automation, triggering on source pushes for immediate feedback. CodeBuild is integrated into the pipeline's build stage to execute tests, ensuring code quality before deployment. CodeDeploy is the ideal service for deploying Lambda functions with automated traffic shifting, and the CodeDeployDefault.LambdaLinear10PercentEvery3Minutes option directly fulfills the requirement for incremental, percentage-based traffic shifting, maximizing developer velocity by automating a complex deployment strategy. The other options are less optimal: The first option incorrectly suggests a post-commit hook triggers CodePipeline after tests pass, which is not a standard CodePipeline trigger mechanism. CodePipeline typically triggers on source changes. The second option introduces unnecessary complexity by using CloudFormation to trigger CodePipeline and doesn't explicitly mention a build/test step within the pipeline. The fourth option relies on a custom, complex, and error-prone solution using AWS CLI, S3 events, and a custom Lambda for deployment and traffic shifting, which is less robust and scalable than using CodeDeploy's built-in capabilities.