Continuous web performance monitoring using Node.js
summary
A company with 300 million users implemented continuous, user-centric web performance monitoring using Node.js by running Lighthouse (headless Chrome) to collect metrics like Time to Interactive and First Meaningful Paint under simulated 3G network and CPU throttling. They run automated per-page Lighthouse jobs that trigger alerts and block CI on SLO regressions, and found that measuring real interactivity (not just full load) reveals different optimization priorities because pages can appear fully loaded while still performing background work.
AI-written summary · about
chapters (7)
- 0:00Start
- 3:26Monitoring the Navigation Timing API
- 8:16Monitoring user-centric metrics
- 11:36Lighthouse!
- 12:47Lighthouse architecture
- 13:34Using Lighthouse programmatically
- 14:36Network throttling
related talks
35mImprove Your Web Application Using Progressive Web Metrics
14mAutomate Performance Monitoring for Micro Frontend Apps
22mUsing Lighthouse in the Real World
16mFrom beginner to expert in web performance
31mAutomated Performance Testing With WebDriver
16mFrom Beginner to Expert in Web Performance
35mWeb Application Performance Testing from Users Perspective
43mDemystifying and leveraging Web Vitals