AWS CodePipeline

AWS CodePipeline is Amazon's managed CI/CD release service for teams on AWS, chaining source, build, test and deploy stages, with usage-based billing.

AWS CodePipeline moves each code change from a source repository through build, test and deployment without a person handing it along. A release is modelled as a pipeline, an ordered set of stages, and each stage holds actions that do one job such as fetching source, running a build or deploying. It is a managed service from Amazon Web Services, part of Amazon. CodePipeline orchestrates the release, while the build and deploy work runs in the services its actions call. Source actions watch a repository in GitHub or AWS CodeCommit. On V2-type pipelines a trigger can filter by branch, file path or Git tag, or start when a pull request is opened, updated or closed. An Amazon EventBridge rule can also start a pipeline on a schedule. Build and test actions hand work to AWS CodeBuild, and other actions release through AWS CodeDeploy, AWS CloudFormation, Amazon ECS, Amazon EKS or AWS Lambda. Each action passes its output artifact to the next as input, so later stages work on exactly what earlier ones produced. An approval action pauses a stage until someone with the right IAM permissions approves or rejects it. CodePipeline is meant for teams whose applications and infrastructure already live on AWS. Platform and DevOps engineers use it to give each service one release path, with an approval in front of production. Access is governed by AWS Identity and Access Management, so the people who may approve a release are the ones the account's policies name. The Commands action, pipeline-level variables and source revision overrides let engineers adjust a run without rebuilding the pipeline. It is bought by usage on an AWS account, next to the services its actions call. A common pipeline starts when a developer pushes to the main branch, builds and tests the change in CodeBuild, then deploys it to a staging stage. A reviewer checks staging before the production stage runs. Infrastructure teams follow the same path with CloudFormation stacks, container teams push images from Amazon ECR to ECS or EKS, and mobile teams add AWS Device Farm tests before a release. For a team already on AWS, it turns a hand-run release checklist into one repeatable pipeline per application.