Observable Web Applications
summary
Argues that client-side measurement is the most important view, showing concrete fixes: move monitoring agents from inside the cloud to the browser (and once to a fallback “box under a desk”) to detect real-user outages; captures errors like stripe checkout is not defined and unexpected token "<" (HTML returned instead of JSON); uses track.js, request metrics and Kibana-style logging; shows p75 load time rising from ~4s to ~10s (95th to ~20s) and conversion dropping from 25% to 18%.
AI-written summary · about
chapters (7)
- 0:00Observable web apps intro
- 10:42Web app architecture
- 17:51Client-side error tracking
- 31:58Performance monitoring basics
- 49:49Usage analytics & funnels
- 55:03Client-side observability limits
- 57:52Summary
related talks
33mRedefining Web Performance
39mGiving you ways to increase, but also explain site performance, to your boss.
20mReal metrics for front-end performance
06mMeasurement is Not Enough: Rise of Performance Analytics
16mCoding Session: Stunning Performance
54mDid we(b development) lose the right direction?
24mVanity performance metrics are worthless
48mHow I learned to stop worrying and love UX metrics