From FTP to Kamal: 20 Years of Deploying Rails

on YouTube ↗26m

" summary

Traces 20 years of Rails deployment from FTP and CGI to Mongrel’s C HTTP parser, FastCGI fragility, Passenger isolating apps inside nginx, Capistrano automating SSH-based releases, Heroku turning ops into a one-liner and Puma/Unicorn changing concurrency models. Claims Docker finally solves Ruby-version drift and Kamal replaces Capistrano with containerized releases. Notes surprising reliability improvements and persistent pain points: process crashes, memory leaks, and the moment most remember as git push heroku alongside Capistrano and Docker and Kamal.

AI-written summary · about

" chapters (10)

1. 0:00Introduction
2. 2:34Web dawn and early deployments
3. 7:25PHP and shared hosting
4. 11:13Rails debut and challenges
5. 15:05Mongrel and HTTP servers
6. 18:15Capistrano and automation
7. 20:41Unicorn, Puma, and servers
8. 22:24Heroku and cloud era
9. 23:31Docker and containers
10. 25:58Summary and call to action

" related talks

• 35mTen Years of Rails DeploymentChristopher Rigor · Rails Pacific 2014
• 35mMy Rails App is Old Enough to Drink: Over Two...John Paul Ashenfelter · Blue Ridge Ruby 2024
• 07mLightning Talk: How Docker Can Change Rails DeploymentsChristopher Rigor · RedDotRuby 2015
• 36mContinuous Deployment with RailsSebastian von Conrad · RubyConf AU 2014
• 34mGuide to Continuous Deployment with RailsKeith Pitt · Red Dot Ruby Conference 2014
• 43mServer Infrastructure for Rails in 2016Richard Lee · Rails Pacific 2016
• 49mRuby Software Continuously Delivered and Exhaustively ExplainedBrian Guthrie, Srushti Ambekallu · RubyConf 2011
• 31mTen Years of Rails TutorialsMichael Hartl · RailsConf 2018