Building pipelines with CodePipeline

Automating Deployments on AWS

Dunieski Otano

Amazon Web Services Solutions Architect

Multi-stage pipelines

Anatomy of a CodePipeline pipeline

• A pipeline is a sequence of stages
• Each stage contains action groups with actions
• Actions produce and consume artifacts (stored in S3)
• The next stage consumes artifacts as input

Connecting source, build, and deploy

• Source stage: CodeCommit detects a push, outputs source artifact
• Build stage: CodeBuild receives source, outputs build artifact
• Deploy stage: CodeDeploy receives build artifact, deploys to target
• Pipeline runs end-to-end on every code push

Manual approval actions

• Manual approval: Human gate between stages
• Pipeline pauses until someone approves or rejects
• Common use: gate between staging and production
• Status shows "Waiting for approval"; supports SNS notifications

Stage transitions and pipeline control

• Enable/disable transitions between stages
  • Disable to pause the pipeline at a specific point
• Retry failed actions without restarting the pipeline
• Release change manually to push a specific revision

 

Pipeline event notifications

• CodePipeline sends events to Amazon EventBridge
• Events include: pipeline started, stage succeeded, stage failed
• Create EventBridge rules to route events to targets
  • SNS topics for email/SMS alerts
  • Lambda functions for custom automation
• Event detail includes pipeline name, stage name, and failure reason

Multi-environment pipeline design

• Single pipeline deploys to multiple environments
• Build once, then promote: Dev -> Staging -> Approval -> Prod
• Same artifact flows through every environment
• What you tested is exactly what you deploy

Let's practice!

Automating Deployments on AWS