Your API is Too Slow!
summary
Claim: most APIs are already fast enough, but when speed matters apply concrete fixes: put traffic behind a CDN and use HTTP/2, deploy read replicas near users, profile in production with ruby-prof and rack-mini-profiler and watch 75th/95th percentiles, push aggregation into SQL instead of heavy Ruby loops, use multi‑layer caching (app memory vs Redis), upgrade gems (observed 10–20% gains), swap JSON parser for OJ, move work to background jobs, and avoid full rewrites — Ruby scaled to 500k requests per minute in one real case.
AI-written summary · about
chapters (7)
- 0:00Is your API fast enough?
- 2:20Defining speed as a feature
- 5:04Network and infrastructure
- 11:20Profiling and measuring
- 14:00Leveraging the database
- 18:54Caching and mutation
- 23:23Optimization and conclusion
related talks
30mOn API Optimization
18mUsing APIs appropriately
29mAccelerate API Innovation: Secrets to Staying Ahead
42mBuilding Better Web APIs with Rails
28mAPI Performance A journey down the rabbit hole
45mAPI on Steroids
24mAPI Optimization Tale: Monitor, Fix and Deploy
28mBetter APIs through Better More Timely Data