Service Extraction at Groupon Scale
summary
Groupon faces real scale: 40+ million active global customers, spikes to 30k requests per minute, a 2M‑line Rails monolith and 2k commits a month that cause 45‑minute test runs and three‑hour deploys. The actual work is extracting an orders service, building a service wall to detect illegal model access, a Route 66 log analyzer to find cruft, greenfield services with a message bus and one‑way background sync, and a manager accessor pattern for merging legacy and international data — all pragmatic tools that prove the approaches.
AI-written summary · about
chapters (10)
- 0:00Start
- 1:52Rails is agile-friendly and that's awesome
- 4:00blame Oprah
- 5:08example: casual coupling
- 7:19it was painful
- 21:38taming COBRAs is serious YPAGNIRN
- 22:29have a strategy
- 23:26envision the whole stack
- 27:08avoid tangled parts has many problems
- 30:09in conclusion...
related talks
22mNitty Gritty Service Building
28mNitty Gritty Service Building
33mServices and Rails: The Shit They Don't Tell You
25mI've Made a Huge Mistake: We Did Services All Wrong
33mServices and Rails: The Sh*t They Don't Tell You
34mSOA and the Monolithic Rails App Anti-Pattern
23mAbstracting a Service - A Case Study in Forward Thinking with SOA
35mLate, Over Budget, & Happy: Our Service Extraction Story