A company deploys an application using EC2 instances and multiple CloudFormation stacks. Developers build and test locally, then upload artifacts and templates to S3; peers review changes before stack updates. The deployment process is error-prone and slow when updating each instance. The company wants to automate deployment but keep a final manual approval before modifying application or resources. Source and templates are already in CodeCommit and a CodeBuild project exists. Which steps meet the requirements? (Choose two.)

The company needs to automate deployments with a manual approval step, using existing CodeCommit and CodeBuild. AWS CodePipeline is the correct service for orchestrating this workflow, allowing integration of CodeBuild, CloudFormation, and CodeDeploy, along with a manual approval action. Creating CloudFormation change sets before approval ensures that the proposed infrastructure changes are reviewed and understood, preventing unexpected modifications. After approval, these change sets can be executed. AWS CodeDeploy is the service designed for automating application deployments to EC2 instances. To use CodeDeploy, an application and a deployment group must be defined, and the CodeDeploy agent must be installed on the target EC2 instances. The incorrect options either miss critical steps (like creating change sets for CloudFormation, or installing the CodeDeploy agent), suggest an incomplete CodeDeploy setup, or propose an order of operations that doesn't align with best practices for reviewing infrastructure changes before deployment. For example, deploying with CodeDeploy before running CloudFormation change sets would mean the application is deployed to potentially un-updated infrastructure.