The Road to continuous deployment
summary
A real migration story: a monolith called San Diego gets strangled by new services behind a proxy, all shipped as Docker containers and built with Jenkins pipelines; deployments use rolling updates, health checks and feature toggles. Achieved under 10 minutes push-to-prod per surface and 50+ deployments per day, page loads fall from ~5–6s to ~0.5s, outages and lead time plummet, 100% PHP coverage gate (later relaxed for JS), and continuous master-only commits plus pair programming drive the change.
AI-written summary · about
chapters (13)
- 0:00Start
- 2:57THE SYSTEM - SAN DIEGO
- 7:02GOALS
- 10:45APPROACH
- 13:26PROXY
- 20:37WHY CONTINUOUS DEPLOYMENT
- 29:38FEATURE TOGGLES, AB TESTS
- 40:19AUTOMATION
- 40:49DEFENSE IN DEPTH
- 44:46CONTINUOUS TESTING
- 46:18PIPELINE AS CODE
- 47:07DOCKERFILE
- 50:48LESSONS LEARNED
related talks
53mThe Road to Continuous Deployment: a Case Study
49mThe road to continuous deployment: a case study
34mEnabling a Continuous Deployment Mindset
43mContinuously deploying a distributed monolith
33mEnabling Microservices
35mPractical PHP Deployment with Jenkins- 35mPractical PHP Deployment with Jenkins
12mContinuous Deployment in a heavily regulated environment