Master LCP, INP, and CLS measurement, debugging, and optimization in production environments.
- LCP Measurement & Optimization
- INP Tracking & Debugging
- CLS Reduction Strategies
- FCP & TTFB Analysis
- View all 6 topics →
Production Engineering Guide
Real-User Monitoring &
Core Web Vitals
Production-focused resources for measuring, analyzing, and optimizing real-user web performance. Field data, p75 baselines, self-hosted pipelines — built for engineers who ship.
Modern web engineering demands more than Lighthouse scores. Real users experience your application across throttled mobile networks, fragmented device ecosystems, and concurrent background processes that synthetic tools simply cannot model. This guide provides the technical foundation for capturing, analyzing, and acting on field data — the only measurement that reflects what your users actually experience.
From instrumenting Core Web Vitals (LCP ≤ 2.5s · INP ≤ 200ms · CLS ≤ 0.1) with the PerformanceObserver API, to designing high-throughput self-hosted beacon pipelines backed by ClickHouse, to correlating p75 latency distributions with conversion funnels — every guide here is production-tested, percentile-anchored, and focused on shipping measurable improvements.
Whether you are a frontend engineer debugging INP spikes in production, a performance architect evaluating OpenTelemetry for RUM, or a technical lead quantifying the ROI of web performance — you will find actionable, engineering-grade resources across all three sections below.
Engineering Guides
Design, build, and deploy self-hosted RUM pipelines with ClickHouse, OpenTelemetry, and Grafana.
- Self-Hosted Beacon Collection
- OpenTelemetry for Web RUM
- RUM Data Sampling Strategies
- Grafana Dashboards for Web Performance
- View all 6 topics →
Define, instrument, and correlate custom performance metrics with conversion rates and business KPIs.
Explore guideCore Web Vitals Thresholds
Assessed at the 75th percentile (p75) over a 28-day rolling window
| Metric | What it Measures | Good | Needs Improvement | Poor |
|---|---|---|---|---|
| LCP Largest Contentful Paint |
Loading performance — time to render the largest visible element | ≤ 2.5s | ≤ 4.0s | > 4.0s |
| INP Interaction to Next Paint |
Responsiveness to all user input throughout the page lifecycle | ≤ 200ms | ≤ 500ms | > 500ms |
| CLS Cumulative Layout Shift |
Visual stability — unexpected element movement during load | ≤ 0.1 | ≤ 0.25 | > 0.25 |