Optimising Continuous Delivery
summary
Claims a working deployment pipeline gives release confidence and must be optimized for speed: aim for commit stage under five minutes (90 seconds ideal) and acceptance under 45 minutes. Shows a GitHub Actions demo built in a day and a half, LMAX pipeline running ~22k UI acceptance tests in under 45 minutes, and a C++ build cut from 9.5 hours to ~12 minutes by faster hardware and parallel builds. Recommends Bazel/Buck, Docker/Kubernetes, versioned pipeline-as-code and treating intermittent tests as failures.
AI-written summary · about
chapters (14)
- 0:00Workshop introduction
- 2:08Core principles
- 6:19Commit cycle strategy
- 11:20Acceptance testing
- 17:01Slow commit stages
- 23:25Failing tests handling
- 31:53Build optimization
- 41:53Static analysis efficiency
- 48:42Commit stage performance
- 51:48Acceptance test rehearsal
- 1:06:06Scaling acceptance testing
- 1:11:04Discipline for fixing failing tests quickly
- 1:16:42Pipeline changes and IaC
- 1:20:14Q&A
related talks
33mContinuous Delivery Pipelines: How to Build Better Software Faster
01h 01mA team's transition to Continuous Delivery
27mOptimizing CI/CD Pipelines for Faster Time to Market
55mThe Rationale for Continuous Delivery
52mThe Future of Continuous Delivery
26mOptimizing Your CI Pipelines
24mCI/CD Best Practices to Deliver Software Faster
44mSimplifying and Scaling Software Delivery