Pipeline Patterns and Antipatterns - Things your Pipeline Should (Not) Do
summary
Shows how to cut pipeline friction with concrete moves: use GitHub Actions YAML or invoke make/just locally, cache build output (target/obj directories) to save minutes — one example saves ~4.5 minutes after a 30s restore — and prefer buildkit/docker buildx remote caches for image layers. Split PR, merge and nightly jobs; publish tooling images and use tooling images with unique git SHA tags to avoid interference; evaluate Dagger or Earthly for reproducible, local-first builds.
AI-written summary · about
chapters (12)
- 0:00Introduction and goal
- 13:36Artifact hoarding
- 14:38Pipeline step ordering
- 17:05Parallelizing pipelines
- 20:23Avoiding tool wrappers
- 23:21Pipeline complexity
- 26:24Local scripting for CI
- 29:18Intelligent caching
- 35:40Container caching
- 38:26Tooling images
- 43:34Build agent sizing
- 48:44Future of pipelines
related talks
27mPipeline Patterns and Antipatterns - Things your Pipeline Should (Not) Do
30mPipeline patterns to aid your next deployment
27mOptimizing CI/CD Pipelines for Faster Time to Market
21mStrategies for Testing Your Pipeline
31mImplementing CI/CD at NGINX -- Lessons Learned Along the Way
39mRemove your Testing Bottleneck with CI/CD Pipelines
32mContinuous Integration and Deployment: Five Lessons Learned
41mContinuous Integration Tips & Tricks