Uploaded byChristopher Read
Continuous Integration, Build Pipelines and Continuous Deployment
This document discusses core concepts and best practices for continuous integration (CI), build pipelines, and deployment. It recommends having a single source code repository, automating builds and testing, publishing the latest build, committing code frequently, building every commit, testing in production environments, keeping builds fast, ensuring all team members can see build status, automating deployment, and making CI and continuous deployment a collaborative effort between developers and system administrators. The goal is to improve quality, time to market, and confidence through practices that provide fast feedback on code changes.
More Related Content
Similar to Continuous Integration, Build Pipelines and Continuous Deployment
Continuous Integration, Build Pipelines and Continuous Deployment
- 1. CI, Pipelines and Deployment Chris.Read@ThoughtWorks.com chris-read.net
- 2. About Me • Unix Admin • Network Admin • Developer • Principal Consultant
- 3. CI Core Practices Photo by teclasorg
- 4. Single Source Repository • Single point of truth • Everyone's code in the same place • NOT a branch per developer • Shared ownership
- 5. Automate Build • Using the IDE is not automating! • Using the IDE does not give you a repeatable build • Use a build tool • Compile, package, test
- 6. Automate Testing • Not just Unit Tests • Failing tests fail the build • Fix it if it’s broken or you’re wasting your time
- 7. Publish Latest Distributable • Make it easy to get the final product • Should only be built once so you deploy what’s been tested • Configuration is separate
- 8.
- 9. Commit More Often • “At least once a day” - aim for at least once an hour • Needs a small unit of work • To commit cleanly you need to update first
- 10. Every Commit Builds • It’s all about fast feedback • Small changes • Less to merge and/or fix
- 11. Test in Production Clone • Detect multi-threaded or cluster issues • Tests system architecture and upgrades • Includes databases! • It’s about moving parts, not capacity
- 12. Keep Build Fast • It really is all about feedback • If things break you find out about it while it’s still fresh in your mind • Keep up with frequent check ins
- 13. Everyone Sees What’s Happening • Reduce time to fix • No excuse to check in on broken build • It’s not about blame, it’s about feedback
- 14. Automate Deployment • Reduce Human Error • Verify you can get it running somewhere other than “my machine” • Test not only the code but your deployment mechanism too • Don’t tie up Sys Admins with boring stuff • Don’t tie up Devs waiting for feedback
- 15. Making CI Work • Can’t be done in isolation • Pick the right tools for the job • It’s not a silver bullet Image by LiminalMike
- 16. Build Pipelines Photo by rickz
- 17.
- 18.
- 19.
- 20. One Click Deploys • Pre-requisite for build pipelines • Reduces deployment time and risk • Makes go live a non-event
- 21. Continuous Deployment Photo by stevendepolo
- 22.
- 23. Summary • CI and Build Pipelines improve • Quality • Time to Market • Confidence • Continuous Deployment is the next step • These practices require Devs and Sys Admins to work closely together
- 24.
- 25. Useful Links • http://www.martinfowler.com/articles/ continuousIntegration.html • http://www.thoughtworks.com/pdfs/ deployment_production_line.pdf • http://saucelabs.com/blog/index.php/2009/09/continuous- deployment%E2%80%94the-video/