Ci/Cd Pipeline Architecture — AWS Solutions Architect Pro (SAP-C02)
CodePipeline Orchestrates; CodeBuild Runs; CodeDeploy Ships
CodePipeline manages the pipeline workflow: source trigger, stage sequencing, and approval gates. CodeBuild runs build and test steps in managed, ephemeral containers with no infrastructure to provision. CodeDeploy handles the deployment mechanics: in-place updates for EC2, blue/green traffic shifting for ECS, and Lambda alias traffic shifting for serverless targets. 'No servers to manage for the build system' specifically identifies CodeBuild, not Jenkins on EC2. Missing the three-way split between orchestration, build execution, and deployment mechanism produces answers that assign the wrong service to each responsibility. SAP-C02 pipeline questions often include Jenkins as a distractor when the operational constraint rules out self-managed build infrastructure.
What This Pattern Tests
The exam presents deployment scenarios and tests strategy selection. CodeDeploy to EC2/ECS supports in-place (rolling), blue/green (two environments, instant cutover/rollback), and canary (shift 10% of traffic, wait, shift remaining). CodePipeline orchestrates the build-test-deploy chain with approval gates. The trap is choosing blue/green for a cost-sensitive deployment (doubles infrastructure) or rolling for a zero-downtime requirement on a single-instance deployment (rolling needs multiple instances). CloudFormation stack updates with change sets vs. SAM deploy for serverless vs. CDK for infrastructure-as-code add another decision layer.
Decision Axis
Risk tolerance (instant rollback vs. gradual validation) and cost constraints (double infrastructure vs. in-place) determine deployment strategy.
Associated Traps
More Top Traps on This Exam
Decision Rules
Whether the pipeline includes CodeDeploy as a deployment execution engine with native ALB health-check integration and automatic rollback capability, versus a CodePipeline + CodeBuild design that automates source-to-artifact promotion but lacks a deployment lifecycle layer and therefore cannot natively evaluate fleet health-check failure rates or trigger rollback without custom scripting.
Domain Coverage
Difficulty Breakdown