Using Lighthouse in the Real World
summary
Shows a concrete Lighthouse workflow: a custom CLI runs 9 critical user journeys hourly, generates Lighthouse JSON, screenshots and HAR, uploads HTML to S3 + DB + Grafana and tracks metrics like LCP/TTFB/TBT over time. Uses Elastic and Chrome UX for field web vitals, enables HTTP/2, and tries fixes from low-hanging image/js optimizations up to architecture changes behind feature flags. Uses hypothesis testing in staging, rolls changes to production behind flags, then measures rollback or removal after one to two weeks.
AI-written summary · about
chapters (18)
- 0:00Leonardo Faria
- 1:47What is Lighthouse?
- 2:46How to use Lighthouse?
- 4:00Performance weighting
- 4:46Core Web Vitals
- 5:03Largest Contentful Paint
- 5:25First Input Delay
- 5:49Cumulative Layout Shift
- 8:23What is a HAR file?
- 9:53Critical User Journeys
- 12:07Automated lab tests alternatives
- 12:47Field data collection options
- 14:43Priorities
- 15:35Low-hanging fruit
- 16:14The Scientific Method
- 16:44Feature Flags
- 17:25Bigger projects: Technical spec
- 17:58Bigger projects: Non-Technical spec
related talks
27mWays of Frontend performance auditing with Lighthouse
35mExploring the Lighthouse
56mContinuous Accessibility: Adding Lighthouse to Your CI/CD Pipeline
20mBringing Lighthouse to Angular
43mHouse of Light
22mYou’re Probably Using Lighthouse Wrong
36mProgrammatically Performant
16mFrom beginner to expert in web performance