Monsters of Web Performance
summary
Finds a 9 MB JS bundle containing a 7 MB SVG flags payload and replaces it with dynamic import splitting to reduce initial download, saving about 700 milliseconds per page render; identifies a 2 MB CSS bundle where only 1% is used and removes unused UI library CSS with webpack bundle analyzer and Coverage to cut render-blocking time; fixes jQuery load HTML parsing that costs 400–3,000 ms, uses parallel small data-fetching bundles, Chrome performance profiler and React DevTools profiler, and sets a 250 ms tab-switch budget with RUM (CloudWatch RUM / Sentry).
AI-written summary · about
chapters (19)
- 0:00Intro
- 2:20The Bumbler
- 2:50The Problem
- 4:09Use Dynamic Import
- 5:17Dont Run Away
- 6:09Out of Control
- 6:52Show Coverage
- 8:16Megabyte CSS
- 9:31Modern Tools
- 11:20Browser Speed
- 12:07Jamstack Pattern
- 13:38Timing
- 15:31Browser Network
- 15:54Creepy Crawlies
- 16:34Infinite Loops
- 17:50Performance Profilers
- 19:07Lab Tests
- 19:20Real User Monitoring
- 20:29Summary
related talks
16mThe Evolution of Web Performance
21mEvolution of Web Performance
26mWeb Performance in modern JavaScript world
41mHow we make Web Apps lightning fast at Microsoft!
28mDecoding JavaScript: Loading Performance —
37mSecret Shortcuts of Loading Web Performance
36mPerformance for the long haul
24mSuperpowers of browser's Web API
