Monitoring for Frontend Performances
Real user monitoring collects performance data from actual visitors' browsers, using the web-vitals library or the PerformanceObserver API and sending results to an analytics endpoint. It shows how metrics vary by device, country and page, and alerts you when a release causes a regression.
Stage: Production · Part of the Frontend Performance roadmap · Last reviewed 19 November 2026
What to learn in Monitoring
- Capture LCP, INP and CLS with the web-vitals library
- Send metrics to an endpoint using navigator.sendBeacon
- Segment results by device, connection and page template
- Report percentiles such as p75 rather than averages
- Set an alert for a metric regression after a deployment
Monitoring interview questions
Why look at the 75th percentile instead of the average?
Averages hide the slow experiences that a tail of users has. Core Web Vitals are assessed at the 75th percentile, so a page passes only if most real visits are good, which better reflects user experience than a mean skewed by fast outliers.
What is the difference between synthetic monitoring and RUM?
Synthetic monitoring runs scripted tests from fixed locations and devices on a schedule, giving consistent trends and early alerts. RUM records what real users experience, covering varied devices and behaviour. They complement each other.
Where Monitoring fits in the Frontend Performance roadmap
Monitoring belongs to the Production stage (advanced level). Keep it fast.
Other units in this stage
Back to the full Frontend Performance roadmap