How Your Deployment Strategy Influences How People Code
summary
Compares four concrete strategies—manual, scheduled (morning), continuous delivery and full continuous deployment—and shows how each changes testing, rollback practices and developer incentives; cites a 2013 Facebook analysis that commits made on deadline show lower quality; gives a team goal of getting a commit from master into production in under three hours; recommends canary deployments (route a subset of users and measure error/CPU/JS exceptions) and feature flags to toggle risky code without a full rollback; contrasts rollback vs fix‑forward tradeoffs.
AI-written summary · about
chapters (7)
- 0:00Start
- 1:07What is a deployment strategy?
- 2:52Types of deployment strategies
- 11:08Deployment scheduling, tooling, and tactics
- 22:17Don't try to make a "build cut"
- 22:52Infrequent deploys hide developer pain points
- 24:20Automate what's reasonable for you - one size does not fit all
related talks
44mI Deploy On Fridays
31mFireside Chat: Continuous Deployment
23mContinuous Deployment - The Last Frontier for Continuous Delivery
26mThe Many Ways to Deploy Continuously
47mI Deploy on Fridays (And Maybe You Should Too)
01h 01mI deploy on Fridays (and maybe you should too)
01h 28mThe fantastic four coding patterns of Continuous Delivery
40mContinuous Delivery Anti-Patterns