Scaling R&D to thousands of people without dedicated QA
summary
Shopify scales R&D to thousands without a dedicated QA by making developers own changes end-to-end, running ~800 deploys per day companywide (about 40 to core), and serving 83,000 requests/sec for 800,000 merchants. Concrete tooling enables this: local env tooling dev plus dev top-hat for exact test snapshots, an orchestrator ship it that merges and rolls out via 5% canaries, and Slack ops via spy with deploy dashboards.
AI-written summary · about
chapters (27)
- 0:00Context
- 3:14QA
- 4:32Tools
- 13:11Application
- 13:41Shipit
- 15:16Chip
- 18:08Emergency mode
- 19:21Canary Stack
- 21:01Merge Queue
- 21:29Chat Ops
- 22:46Dashboard
- 24:45Production
- 25:14Platform Health
- 25:59Shipping
- 26:18Dev ATC
- 27:36Tools and Process
- 28:28What is culture
- 29:19Trust battery
- 29:58Constant learning
- 30:33Open internally
- 31:14We thrive on change
- 31:52Merchants obsessed
- 32:54Engineering at Shopify
- 33:28signposts
- 34:00reduce friction
- 34:40long didnt read
- 38:36closing thoughts
related talks
34mShitlist-driven development and other tricks for working on large codebases
28mKeeping Developers Happy with a Fast CI
17mHow We Deploy Shopify
34mShitlist-driven development and other tricks for working on large codebases
40mHow Shopify Scales Rails
21mTesting Rails at Scale
01h 09mScaling up with Embedded Testing
24mContinuous Deployment at Scale
