1.3Automated Testing and Artifact Management

Embed quality gates in the pipeline so only tested/approved changes proceed. Version and reuse build artifacts.

1.3.1Quality gates and artifacts

• Automated tests: run in Build/Test stages; stop the pipeline on failure to prevent bad deploys.
• Manual approval action: insert a human gate before production deploy.
• Artifacts: version container images in ECR and libraries/packages in CodeArtifact.
• CodeGuru: assist quality with code review (Reviewer) and performance profiling (Profiler).

Quality gates design "at which stage, what you verify, and how you record it." CodeBuild test reports ingest JUnit/TestNG results to visualize pass rates and trends in the console. Tests escalate unit → integration → E2E, and catching defects earlier (shift-left) is cheaper. A manual approval action integrates with SNS to notify approvers and records who approved and when. Mismatching artifacts is a frequent exam theme: container images = Amazon ECR (image scanning, lifecycle policies, immutable tags), language packages = CodeArtifact, and build outputs (zip, etc.) = S3 are clearly separated. ECR image scanning (Basic/Enhanced) embeds vulnerability detection in the pipeline, with Enhanced integrating Amazon Inspector. CodeGuru Reviewer uses ML to flag code defects and security issues (e.g., hardcoded secrets), while CodeGuru Profiler pinpoints runtime CPU/memory hotspots. Combining these to enforce "can’t proceed unless it passes" gates at each stage is the core of DevOps quality assurance.

1.3.2Services that complement the SDLC

Beyond the core CI/CD services (CodePipeline/CodeBuild/CodeDeploy), in-scope exam topics include services that automate container deployment, support SDLC peripheral operations, and standardize toolchain bootstrapping.

1.3.2.1Container deployment automation

App Runner is a fully managed service that automatically builds, deploys, and scales web services and APIs from a source-code repository or container image. AWS fully manages load balancing, auto-scaling, and TLS certificate provisioning, significantly reducing operational overhead compared to ECS/Fargate. In a DevOps context, the selection criterion is "minimize infrastructure management when publishing a web container"; use ECS when fine-grained network control or integration with an existing ECS platform is required.

App2Container (A2C) analyzes existing Java/.NET applications running on a server and automatically generates a Dockerfile, ECS/EKS task definitions, CloudFormation templates, and a CI/CD pipeline scaffold. Its DevOps role is automating legacy-app modernization, reducing the cost and error rate of manual containerization. Select it during the phase of migrating existing applications to containers.

AWS Copilot is a CLI for building, deploying, and operating containerized applications on ECS, Fargate, and App Runner. Declare the service type (e.g., Load Balanced Web Service, Backend Service) and environments (test, prod, etc.), and it automatically generates CloudFormation stacks, a pipeline, VPC, and logging configuration. In a DevOps context, use it when you want to stand up a container environment and CI/CD pipeline declaratively with minimal steps. Note: it is unrelated to GitHub Copilot.

1.3.2.2SDLC support and reuse tools

Serverless Application Repository (SAR) is an AWS-managed repository for searching, deploying, and publishing serverless applications defined as SAM templates. Reusable serverless components (auth flows, image processors, notification handlers, etc.) can be deployed in near-one-click, and the repository is also used to share internal common components within an organization.

CloudShell is an authenticated shell environment available directly in the AWS Management Console from a browser. The AWS CLI, common runtimes (Python, Node.js, etc.), and appropriate permissions are pre-configured, with 1 GiB of persistent storage. In a DevOps context, use it when you want to run operational commands or scripts immediately without storing credentials locally, or when you need to start troubleshooting instantly.

AWS CLI is a tool for operating AWS services from the command line, embedded in shell scripts, CI pipelines, and automation scripts. AWS SDK (e.g., boto3 for Python) is the official library for calling AWS APIs from each language, providing built-in authentication resolution, automatic retry, and pagination. Both are essential foundations for SDLC automation scripts and are widely used in custom pipeline steps and Lambda handlers.

Note: AWS CodeStar reached end of life on July 31, 2024 and no new projects can be created (the following is for conceptual understanding). CodeStar used project templates and an integrated dashboard to launch CodeCommit, CodeBuild, CodeDeploy, and CodePipeline together, standardizing CI/CD toolchain bootstrapping, and it also provided team-member access management and issue-tracking integration. Today, to stand up a toolchain quickly you configure CodePipeline, CodeBuild, and CodeConnections (or CodeCatalyst) directly.

1.3.3Section summary

• Quality gates = automated tests + manual approval action
• Artifacts = ECR (images) / CodeArtifact (packages)
• SDLC support = App Runner / App2Container / Copilot / SAR / CloudShell / CLI & SDK / CodeStar