A company is using AWS CodeDeploy to automate software deployment. The deployment must meet these requirements:
• A number of instances must be available to serve traffic during the deployment. Traffic must be balanced across those instances, and the instances must automatically heal in the event of failure. • A new fleet of instances must be launched for deploying a new revision automatically, with no manual provisioning.
• Traffic must be rerouted to the new environment to half of the new instances at a time. The deployment should succeed if traffic is rerouted to at least half of the instances: otherwise, it should fail.
• Before routing traffic to the new fleet of instances, the temporary files generated during the deployment process must be deleted.
• At the end of a successful deployment, the original instances in the deployment group must be deleted immediately to reduce costs.
How can a DevOps engineer meet these requirements?