Optimizing Your CI Pipelines
summary
A practical playbook for speeding GitLab CI pipelines: use multi-stage Docker builds or push one built image to the registry and pull in subsequent jobs to avoid rebuilding, watch that runner provisioning alone cost ~54 seconds, parallelize tests with pytest-xdist or split across runners to cut wall time from ~6 to ~5 minutes but double minutes billed, use DAGs or child pipelines to avoid unnecessary waits, enable interruptible: true and cache policies to stop wasted work and speed installs.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 4:50Docker
- 11:58Directed Acyclic Graph
- 14:14Downstream Pipelines
- 16:13Parallelizing Tests
- 17:51Interrupt Pipelines
- 19:12Random Tips and Tricks
- 22:37Conclusion
related talks
27mOptimizing CI/CD Pipelines for Faster Time to Market
36mPractical Tips for Build Optimization for GitLab and Docker Users
21mBuild your JS Pipeline in Incremental Fashion
22mEmbedded Software Development Efficiency Improvement with Docker and Kubernetes
01h 10mCreating efficient, convenient and reliable CI
01h 02mContinuous Integration: Tips & Tricks
02h 47mPipes on Pipelines
01h 26mOptimising Continuous Delivery