AWS Lambda performance tuning
summary
An actual walkthrough shows a Node.js API using Aurora/Postgres + RDS Proxy, Elasticache (Redis) and API Gateway, with live load tests and tooling: Power Tools, Power Tuning and bundlers. Swapping SQL libraries to postgress.js shrinks bundle size from ~1.6MB to ~40KB, ARMs (Graviton) cut ~20% latency, Redis caching saves ~84ms, and configuration tuning yields ~56% fewer cold starts and up to 95% faster warm responses (P99 drops from >2s to sub‑second). The sample stack runs and those gains replicate in tests.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 1:48How Lambda works
- 5:39Unoptimized Lambda API example
- 9:50Performance tips
- 14:07Load testing with Artillery
- 18:51Caching and ARM optimizations
- 22:44Parameter caching with extensions
- 26:23Low latency runtime (LRT)
- 29:14Key takeaways
- 30:00Q&A
related talks
28mAccelerate with confidence through AWS Distributed Load Testing
42mBest practices for serverless developers
29mEvolve your web application for scale
26mBuild a performance-optimized GraphQL API
13mOptimizing Backend API Performance
30mCaching, Payloads, and Other Dark Arts
25mBest practices for advertising and marketing technology on AWS
53mMastering Node.js Performance