Fast & Functional
summary
Builds a minimal Scala microservice that aggregates YouTube-video likes/views, models side effects as small algebras returning an abstract F (higher-kinded effect), and swaps implementations (HTTP client, logger, cache, server) without touching business logic. Uses cats IO, runs load with wrk and profiles with async-profiler to read flame graphs. Measured baseline ~273 RPS (30s); swapping logger and client plus adding a cache pushes throughput into the tens of thousands RPS (∼31k–40k RPS) and exposes remaining HTTP/IO bottlenecks.
AI-written summary · about
chapters (10)
- 0:00Introduction to talk
- 2:30Functional programming approach
- 9:18Scala code walkthrough
- 23:03Plug-and-play architecture
- 25:34Performance testing setup
- 26:18Analyzing flame graphs
- 30:38Optimizing with interpreters
- 36:19Final performance results
- 41:02Key takeaways summary
- 44:33Q&A session
related talks
36mFreestyle: A framework for pure functional apps, libs and microservices
48mHigh performance Scala systems without tears
32mProduction-ready direct-style Scala
24mHigh Performance Serverless Functions in Scala
51mCats Effect: The IO Monad for Scala
50mA FUNCTIONAL SCALA STACK FOR 2020
41mBuilding Low Latency Microservices & Monoliths in Java
49mFunctional programming in practice, with all its consequences