Pipeline Patterns and Antipatterns - Things your Pipeline Should (Not) Do
summary
Concrete patterns and numbers: run the most-failing tests earlier so failures surface in ~12 minutes instead of 24, split backend/frontend and reuse build artifacts to save ~8 minutes, and cache build outputs (2 GB cache, ~30 s transfer) to cut ~4.5 minutes from compile+lint. Use path filters to skip irrelevant jobs, prefer calling CLI tools over opaque wrapper tasks, publish tooling images tagged by commit to avoid interference, and compose small reusable pipeline snippets instead of a 5,000-line Jenkins monolith.
AI-written summary · about
chapters (13)
- 0:00Introduction and goals
- 1:09Pipeline objectives
- 3:11Anti-pattern: The ritual
- 6:40Conditional pipeline steps
- 9:47Ordering and failing fast
- 12:24Parallelizing pipelines
- 16:55Avoiding gift-wrapped tasks
- 20:37Managing pipeline complexity
- 23:54Scripted pipelines
- 26:27Caching strategies
- 34:26Interference avoidance
- 36:11Sizing build agents
- 40:06Conclusion and recap
related talks
27mPipeline Patterns and Antipatterns - Things your Pipeline Should (Not) Do
51mCI/CD Patterns and Antipatterns - Things your Pipeline Should (Not) Do
54mBreaking Down Your Build: Architectural Patterns For A More Efficient Pipeline
02mTechniques to Speed Up your Build Pipeline for Faster Feedback
12mIntegrated Pipelines
25mHow NOT to Build a Pipeline
18mDesign Patterns For Efficient DevOps Processes
47mCI/CD Agility and Controlling Pipeline Sprawl
