CodeDeploy¶

Overview¶

• Each EC2/On-Prem machine needs to be running the CodeDeploy agent.
• Agents poll CodeDeploy for changes.
• CodeDeploy sends the appspec.yml file to the agent.
• The application is pulled from GitHub or S3.
• EC2 will run deployment instructions
• CodeDeploy Agent reports the success/failure of the deployment.
• EC2 instances are grouped by deployment group (dev, test, prod etc).
• Can integrate with CodePipeline, and use CodePipeline artifacts.
• Can use existing tools, applications.
• Integrated with auto-scaling.
• Supports Blue/Green deployments (EC2 only).
• Supports Lambda deployments.
• Doesn't provision resources.

Main Components¶

• Application: Unique name
• Compute platform: EC2, On-Premise (hosts need to be tagged) or Lambda
• Deployment Configuration: Deployment rules for success/failure
• Deployment Group: group tagged instances.
• Deployment Type: In-place, or Blue/Green.
• IAM instance profile: To give EC2 permissions to pull from S3 or GitHub.
• Application Revision: app code + appspec.yml file.
• Service Role: Role for CodeDeply to perform what it needs.
• Target Revision: Target application version.

AppSpec.yml¶

• File section: Where to get the source files from S3/GitHub.
• Hooks: Instructions on how to deploy the new version (have timeouts). The order is:
• ApplicationStop
• DownloadBundle
• BeforeInstall
• AfterInstall
• ApplicationStart
• ValidateService
• BeforeAllowTraffic
• AllowTraffic
• AfterAllowTraffic

Example¶

version: 0.0
os: linux
files:
  - source: /index.html
    destination: /var/www/html/
hooks:
  BeforeInstall:
    - location: scripts/install_dependencies
      timeout: 300
      runas: root
    - location: scripts/start_server
      timeout: 300
      runsas: root
  ApplicationStop:
    - location: scripts/stop_server
      timeout: 300
      runsas: root

Deployment Configuration¶

• One instance at a time. Stop if an instance fails.
• Deploy to 50% of instances.
• All at once. Good for dev.
• Custom: min healthy hosts etc.

Deployment Failures¶

• Instances stay in "Failed" state.
• New deployments will be deployed to failed state instances first.
• Rollback by re-deploying the old deployment or enable automated rollback.

Deployment Targets¶

• Set of EC2 instances with tags
• Directly to an ASG
• Mix of ASG/tags to build deployment segments.
• Customised scripts with DEPLOYMENT_GROUP_NAME environment variables (if in dev, foo, else if in prod, bar).

EC2¶

• Define how to deploy the application using appspec.yml
• Will do in-place updates/blue-green deployment to EC2 instances.

ASG¶

In-Place Updates¶

• Update the existing EC2 instances.
• New instances created by an ASG will also get the automated deployments.

Blue/Green Deployments¶

• A new ASG will be created with identical settings to the existing ASG.
• Can choose how long to keep the old instances.
• Must be using an ELB.
• Full control over deployment steps.

Rollbacks¶

• Can specify automated rollbacks.
• When deployment fails.
• When CloudWatch alarm is triggered.
• Disable rollbacks completely.
• The last known good revision is re-deployed as a new deployment, with a new version id.