Continuous Delivery using AWS CodePipeline, AWS Lambda & AWS ElasticBeanstalk
8 likes1,826 views
The document discusses AWS CodePipeline, a continuous delivery service for streamlined application updates, highlighting its features such as integration with third-party tools and customizable workflows. It details the release process phases, terminology, and provides demos for various deployment scenarios and tips for effective usage. Additionally, it includes suggestions for future enhancements to the service.
Continuous Delivery using AWS CodePipeline, AWS Lambda & AWS ElasticBeanstalk
- 1. Continuous Delivery using CodePipeline, Lambda & ElasticBeanstalk THOMAS SHAW AUGUST 2016 @thomasshaw [email protected]
- 2. Thomas Shaw - August 2016 Slide 2 What am I talking about • What is AWS CodePipeline? • Where does CodePipeline fit in to the product mix? • Code Pipeline Terminology • Deploy Processes • Demo – Basic pipeline (S3 > ElasticBeanstalk) • Demo – Pipeline (Bitbucket > S3 > Lambda > Multi-region ElasticBeanstalk) • Tips, tricks & problems • Feature requests
- 3. Thomas Shaw - August 2016 Slide 3 AWS CodePipeline • Continuous delivery service for fast and reliable application updates • Model and visualise and control your software release process • Builds, tests, and deploys your code every time there is a code change • Integrates with third-party tools and AWS products • Configurable workflow • Easy to integrate • Improved quality • Rapid delivery • Building blocks
- 4. Thomas Shaw - August 2016 Slide 4 Why use it? • Allows powerful modeling of your release processes • Serial and parallel execution • Easily add a new step to your process • Pause for manual approvals • Multiple deployment steps
- 5. Thomas Shaw - August 2016 Slide 5 AWS CodePipeline Announced - November 2014 Public - July 2015 Available in 3 regions • US East (N. Virginia) • US West (Oregon) • EU (Ireland) NOT CURRENTLY AVAILABLE IN AUSTRALIA $1 per active* pipeline per month https://aws.amazon.com/codepipeline/
- 6. Thomas Shaw - August 2016 Slide 6 Release Process Phases
- 7. Thomas Shaw - August 2016 Slide 7 AWS Product Integrations
- 8. Thomas Shaw - August 2016 Slide 8 AWS Service Integrations 6 “Actions” • Approval • Source • Build • Test • Deploy • Invoke
- 9. Thomas Shaw - August 2016 Slide 9 3rd Party Integrations
- 10. Thomas Shaw - August 2016 Slide 10 AWS CodePipeline
- 11. Thomas Shaw - August 2016 Slide 11 CodePipeline Terminology Pipeline Transition Pipeline name Stage Action
- 12. Thomas Shaw - August 2016 Slide 12 CodePipeline Terminology Parallel Actions
- 13. Thomas Shaw - August 2016 Slide 13 CodePipeline Terminology Sequential Actions
- 14. Thomas Shaw - August 2016 Slide 14 NEW Manual Approval – Aug16 You can add a manual approval action where you want the pipeline to stop running until someone approves or rejects the revision. • Pipeline will stop executing when it has reached the action • Pipeline execution resumes only when the action has been approved • Approval action managed with AWS IAM permissions • Pipeline can notify users via SNS – email, SMS, webhooks etc
- 15. Thomas Shaw - August 2016 Slide 15 Why I tried CodePipeline • 3 weeks of experimentation. • Outgrew our existing solution using BitBucket Pipelines • Cross region deployment all at once • Need visualisation & GUI interface • Add “gated” stages • 5 min setup • Increase developer productivity • Lambda functions are very useful
- 16. Thomas Shaw - August 2016 Slide 16 Still in BETA?
- 17. Thomas Shaw - August 2016 Slide 17 Demos • Build a basic pipeline • S3 > ElasticBeanstalk • Add “Approval” stage • Deep dive into CodePipeline > Lambda JSON payload • BitBucket > S3 > Lambda > ElasticBeanstalk (1 env) • BitBucket > S3 > Approval stage > Lambda > ElasticBeanstalk (4 env, 2 regions)
- 18. Thomas Shaw - August 2016 Slide 18 Failed Deploy
- 19. Thomas Shaw - August 2016 Slide 19 Lambda debug
- 20. Thomas Shaw - August 2016 Slide 20 Tips & Tricks • S3 source bucket needs to be in the same region as CodePipeline • S3 bucket – turn on versioning • Poll time ~30sec – 1min • Input/Output artifact name truncation ~10 characters • Lambda log files are great for debug • Understand the CodePipeline > Lambda JSON payload • BitBucket > S3 (using BitBucket Pipelines) • Console delays
- 21. Thomas Shaw - August 2016 Slide 21 Feature Suggestions • Native Bitbucket integration • Ability to pass more environment variables through the stage/actions • Change Poll time • Pipeline/Stage History • Version control • S3 as a Deploy location • Cross region support • Artifact extensions • Mobile app • Documentation
- 22. Thomas Shaw - August 2016 Slide 22 Questions?