How Shopify Scales Rails
summary
Shopify runs Rails at 50k requests per minute with ~45 ms storefront responses and handles spikes to ~200k rpm during flash sales, scaling from a single server to 33 app servers and 1,172 Unicorn workers. The stack uses Ruby 1.9.3, Rails 3.2, Percona MySQL, Redis, memcache, New Relic and Splunk. Built tools include cachable (controller/page memcache caching), identity_cache (memcache-backed model cache) and conan (production load tester); backgrounding (Resque) and DB tuning cut latency and removed contention.
AI-written summary · about
chapters (20)
- 0:00Intro
- 1:23The Stack
- 5:16Looking Back, to Look Ahead
- 10:50Know The System
- 12:03Chive Flash Sale
- 13:27Measure ALL THE THINGS
- 15:18New Relic
- 15:34Splunk
- 20:02Caching 4045
- 21:44Identity Cache
- 24:13Delayed Job
- 24:50Resque
- 26:46Background Payment Processing
- 27:57Redis
- 29:11MySQL Hardware
- 29:58MySQL Query Optimization
- 31:27MySQL Tuning
- 34:38after_commit
- 36:43Services
- 37:22Imagery
