The fantastic four coding patterns of Continuous Delivery
summary
Four concrete patterns enable faster, safer releases: keep work on mainline and ship weekly using track-based development; commit new logic behind defaults so it ships dormant as latent live code and gets exercised by tests; switch behavior with a simple feature toggle while duplicating and testing the new schema; and deploy a binary that auto-detects storage/API version as a forward-compatible interim version. Ferrari teams use telemetry replay, quick DB copy rollbacks (minutes–hours), and automated version checks to catch breaking changes early and keep races running.
AI-written summary · about
chapters (19)
- 0:00Intro
- 6:52Challenges
- 10:06Sample Code
- 12:55Questions
- 16:37Whats next
- 18:07Simulation
- 26:16Trackside
- 27:13Replay
- 31:17Data Barrier
- 34:25Bugs
- 35:53Overtaking
- 36:32Eliminate
- 37:37Latin to Light
- 38:30One more week
- 39:24The next step
- 39:55Feature toggle
- 44:38New version
- 48:28Backwardcompatible breaking changes
- 56:28Forwardcompatible interim version
related talks
40mSustainable Continuous Delivery: Patterns for rapid and reliable Releases
51mBetter software, faster: principles of Continuous Delivery and DevOps
40mContinuous Delivery Anti-Patterns
32mHow Your Deployment Strategy Influences How People Code
59mThe continuous culture
27mDevOps Patterns and Antipatterns for Continuous Software Updates
33mContinuous Delivery - What does good really look like?
46mSix small steps towards continuous delivery