Design AWS Microservices CI/CD Pipeline

My engineering team runs a suite of container-based microservices on AWS, but our current release process is still semi-manual. I want a clean, opinionated CI/CD architecture that begins with every pull request and ends with an automatically promoted production deployment, all driven by GitHub Actions.

Scope of work
• Map out an end-to-end pipeline that covers build, test, security scanning, artifact versioning, staging, and blue/green or canary production rollout.
• Express the infrastructure in code (Terraform or AWS CDK are both fine) so environments can be recreated at will.
• Keep everything native to AWS where it adds value—e.g., ECR for images, S3 for artifacts, CloudWatch for logs—yet let GitHub Actions remain the single orchestration layer.
• Document the workflow clearly so new developers can ship a change without hand-holding.

Acceptance criteria
• A GitHub repository containing reusable Actions workflows per service.
• Infrastructure code that spins up all required AWS resources in one command.
• Successful end-to-end demonstration: a sample microservice merged to main, automatically tested, container pushed to ECR, and deployed with zero downtime to the target cluster.
• Markdown runbook outlining rollback and troubleshooting steps.

If another tool (Jenkins, GitLab CI/CD, or AWS CodePipeline) would materially improve a step, call it out in your proposal; otherwise the default stays GitHub Actions. I value clear documentation, modular code, and pipelines that fail fast and visibly.