Ruby at Scale
summary
Ruby powers a 40M‑line monolith with 500+ engineers and ships features continuously (about 20 deploys a day), while an API example serves 70,000 requests/minute with p50≈25ms and p90≈70ms. Teams use codeowners/service_owners, Packwerk modularization, feature flags, review‑app labs and monitoring (Datadog, Sentry, AppSignal). Performance wins come from OJ for JSON, YJIT and tuned thread counts, rubocop‑performance rules, and incremental DB and caching fixes.
AI-written summary · about
chapters (9)
- 0:00Introduction to Ruby at Scale
- 5:57Does Ruby scale?
- 11:55Building a performant Rails app
- 22:54Ruby in large teams
- 32:10Speed of development and releases
- 44:57Static typing in Ruby
- 53:17GitHub’s tech stack evolution
- 56:12Testing and refactoring in Ruby
- 1:00:52Q&A: Defining scale
related talks
32mBuilding HUGE Web Apps: Rails at 1,000,000 Lines of Code
31mDavid and Goliath, Ruby and the Enterprise
43mReal World Ruby Performance at Scale
28mBringing Your Rails Monolith Along As The Business Grows
32m2000 engineers, 2 millions lines of code: the history of a Rails monolith
14mDatabase Performance at Scale for RoR Applications
01h 00mRuby Committers vs the World
27mScaling Ruby Applications: Challenges, Solutions and Best Practices