Continuous Deployment with Rails
summary
Argues for automatically deploying master on every merge and shows a working pipeline: CI runs Cane, RuboCop and fast tests, buildbox deploys to staging then production, smoke tests verify success, and PagerDuty/Bugsnag/New Relic monitor it. Uses feature flags with rollout gem or flip so changes are hidden until enabled. Handles zero‑downtime with Heroku preboot and tackles migrations strategies. Reports moving from ~1.9 to 5 deploys a day with much smaller, safer changes and a demo social‑button feature that actually ships behind a flag.
AI-written summary · about
chapters (13)
- 0:00Start
- 2:26Pin Payments
- 10:06HOW DO YOU DEPLOY IT WITH ZERO DOWNTIME?
- 10:20HOW WILL WE HANDLE DATABASE MIGRATIONS?
- 10:39Reviewing Code GITHUB PULL REQUESTS
- 15:43BONUS POINTS: THE SAME TESTS RUN AGAINST REMOTE SERVER & LOCAL
- 16:50Zeno Downtime Deploys DEPLOYING NEW CODE WITHOUT MAINTENANCE PAGE
- 19:20EASY WAY: MAINTENANCE PAGE
- 19:45HARD WAY: DEPLOY MIGRATIONS FIRST
- 20:16Database Migrations BEWARE OF DATABASE LOCKING
- 25:28TIME TO MIX
- 26:227 git push heroku master 8 rake smoke:production
- 30:02Merge pull request
related talks
34mGuide to Continuous Deployment with Rails
30mKeith and Mario's Guide to Continuous Deployment with Rails
34mStrong Practices for Rails Applications Continuous Delivery
46mContinuous Deployment Workflows
01h 12mPitfalls of Continuous Deployment
49mRuby Software Continuously Delivered and Exhaustively Explained
42mCONTINUOUS DELIVERY FOR RUBYISTS
35mCI/CD and devops with Ruby and Rails
