On API Optimization

on YouTube ↗30m

" summary

A live API goes from ~5000 ms average to sub‑1s by following a strict workflow: get a baseline, profile with New Relic and a local object profiler (stackprof style), inspect DB with a SQL query tracer, find one application method repeatedly calling an external Google currency conversion, and replace it with a cached lookup and a single DB result. The fix includes added tests, load checks and regression monitoring and measurably cuts median and p95 latencies to acceptable levels.

AI-written summary · about

" chapters (13)

1. 0:00Introduction and speaker background
2. 1:14What is API optimization
3. 3:09Understanding API structure and bottlenecks
4. 5:18Finding the root cause of API slowness
5. 7:30Tools for profiling and monitoring APIs
6. 10:45Profiling with New Relic and other tools
7. 12:54Common performance pitfalls and quick wins
8. 15:24Avoiding major changes for minor gains
9. 16:50Library performance and third-party dependencies
10. 19:52Testing in production and pre-production
11. 21:23Automating performance checks
12. 23:15Case study: Currency conversion API
13. 26:15Summary and key takeaways

" related talks

• 29mYour API is Too Slow!Grzegorz Witek · RedDotRubyConf 2016
• 43mAPI Optimization Tale: Monitor, Fix and Deploy (on Friday)Maciej Rząsa · Build Stuff 2022
• 30mRails Performance monitoring: A primer for developersRishi Jain · RubyConf India 2023
• 26mRails Performance Monitoring 101: A Primer for DevelopersRishi Jain · RubyConf TH 2023
• 24mAPI Optimization Tale: Monitor, Fix and DeployMaciej Rząsa · GopherCon Europe 2023
• 17mHands-On Experience in API Performance Testing | What is API TestingVishal Parmar · TestFlix 2023
• 29mAccelerate API Innovation: Secrets to Staying AheadAnna Daugherty · apidays London 2023
• 31mHigh performance APIs in Ruby using ActiveRecord and GoliathDan Kozlowski · RailsConf 2015