Perf Stories
summary
Case studies and concrete wins matter: code changes cut servers to one third while doubling throughput and halving latency, Pino logs 10,000 messages in ~250ms (100ms in extreme mode), 0x generates flame graphs from macOS/Linux stacks, autocannon saturates servers ~10% more than wrk/ab and scripts full CI benchmarks; discovered 25% gains by manual JSON building, 20–25% by using epoch ms instead of toISOString, and ~60% faster buffer writes after string flattening.
AI-written summary · about
chapters (13)
- 0:00Start
- 1:23Flashback: January 2016 The Frame Story
- 2:23Sensory Detail
- 2:56The network card... became the bottleneck!.
- 5:07Moral: Comment Optimized Code
- 7:16NET-A-PORTER
- 15:35Fighting Red Herrings
- 23:47String Theory'
- 25:01Killing Time
- 26:08Waiting for Closure
- 30:45Rigging
- 33:15Automated Reports
- 33:44Epilogue A Prophecy
related talks
36mMake It Fast
33mReaching Ludicrous Speed
48mFantastic Performance and Where to Find It
46mReal-world HTTP benchmarking
01h 03mThe performance loop—A practical guide to profiling and benchmarking
38mPerformance in the Wild
47mHigh perf architecture at Stack Overflow
46m10 secrets to improve Javascript Performance
