Keeping Your CI–CD Pipeline as Fast as it Needs to Be
summary
Shows how to shrink CI/CD latency by breaking a monolith into modules, publishing every commit as a new version and wiring builds with maven snapshots and TeamCity snapshot dependencies (or Jenkins equivalents) to only rebuild what changed. Explains hacks to avoid infinite triggers and needless rebuilds: auto-bump versions, commit with a CI user and run a fast “skip build” profile when appropriate. Introduces metrics—change rate (example 28%), build time, and weighted impact time—and a prototype tool build-hospice that visualizes hot spots to guide where to split next.
AI-written summary · about
chapters (28)
- 0:00Introduction
- 4:50Talk Outline
- 9:05Why is size a problem
- 19:33What is Maven
- 21:36The problem with Maven
- 25:16The solution
- 27:00Changing the version number
- 28:33Building
- 29:52Scaling
- 30:06Splitting it up
- 33:35Snapshot dependencies
- 34:29Snappy dependencies
- 34:57Does anyone ever have to rebuild all versions
- 35:29Config file
- 36:00Updating dependencies
- 36:52Template process
- 37:25Do you have anything similar
- 37:42The Infinite Trigger
- 39:24The Pending Change
- 40:07Unnecessary Rebuilds
- 41:41Trick Density
- 42:00Double Agent
- 42:26Skip Build
- 43:32Fake Build Example
- 45:22Waiting City
- 48:17Explanation
- 58:25Change the script
- 58:40Change the config file
related talks
24mCI/CD Best Practices to Deliver Software Faster
41mContinuous Integration Tips & Tricks
27mOptimizing CI/CD Pipelines for Faster Time to Market
01h 10mCreating efficient, convenient and reliable CI
02mTechniques to Speed Up your Build Pipeline for Faster Feedback
48mDemystifying CI/CD
51mBetter software, faster: principles of Continuous Delivery and DevOps
32mPractical Tips & Tricks for CI/CD Success