Performance Monitoring Tools
Performance tools split into two kinds that answer different questions: lab tools that test a page under controlled conditions, and field tools that report what real users actually experience. You need both, because a page can score well in the lab and still be slow for users on real devices and networks. This annotated directory groups the dependable options by what they measure and when to use each.
The recurring principle is to measure before and after every change, so optimization targets real numbers.
What you'll learn
Lab auditing tools
Lab tools run a page under fixed conditions, giving repeatable scores for comparison.
Controlled, repeatable scores
- Lighthouse · audits performance, accessibility, best practices, and SEO, available in DevTools and on the command line, the standard first check for a page.
- PageSpeed Insights · runs Lighthouse in the cloud and adds real-user field data alongside the lab score, so one report shows both views.
Real-user monitoring
Field tools capture performance from actual visitors, which lab tests cannot.
What users really experience
- The web-vitals library · a small script that measures Core Web Vitals in the browser and reports them to your analytics, the do-it-yourself path to real-user data, matching the per-route approach in the JavaScript SEO example.
- Commercial RUM platforms · hosted services that collect and chart field metrics over time, worth it when manual collection becomes a burden.
Deep network testing
For detailed, configurable lab testing, a dedicated tool goes further than a single score.
Testing across conditions
- WebPageTest · runs tests from real devices and locations with detailed waterfalls, filmstrips, and connection options, the tool for understanding exactly how a page loads.
- DevTools network throttling · the built-in option for quickly reproducing a slow connection during development, covered in the browser DevTools guide.
The Chrome UX Report
Aggregated field data for millions of sites comes from one public dataset.
Real-world data at scale
- The Chrome User Experience Report · real-user metrics gathered from Chrome users, the source behind the field data in PageSpeed Insights and Search Console, useful for benchmarking against real-world performance.
- The Search Console Core Web Vitals report · presents this field data grouped by your own URLs, showing which page types need work.
Bundle analysis
Much frontend slowness is simply too much JavaScript, which bundle tools expose.
Finding what bloats the bundle
- Bundle analyzers · visualise the size of each module in your built output, making it obvious which dependency is responsible for a large download and worth replacing or removing.
- The DevTools Coverage panel · complements analyzers by showing how much of the shipped code actually runs, exposing dead code to cut.
Continuous monitoring
A one-off audit drifts, so performance is best watched automatically.
Catching regressions early
- Lighthouse CI · runs Lighthouse automatically on each change and can fail a build when scores drop, which stops performance regressions before they ship.
- Scheduled field monitoring · tracking real-user vitals over time reveals slow degradations that a single test never would, the ongoing habit behind the performance roadmap.
Frequently Asked Questions
What is the difference between lab and field performance data?
Lab data comes from testing a page under controlled conditions, giving repeatable scores. Field data comes from real users on real devices and networks. A page can score well in the lab yet be slow for users, so you need both.
Which performance tool should I start with?
Lighthouse, available in DevTools and PageSpeed Insights, is the standard first check. PageSpeed Insights is especially useful because it shows both a lab score and real-user field data in one report.
How do I measure performance for real users?
Use the web-vitals library to measure Core Web Vitals in the browser and report them to your analytics, or a commercial real-user monitoring platform. The Chrome UX Report also provides aggregated field data.
How do I stop performance from regressing?
Run Lighthouse CI on each change so a build can fail when scores drop, and monitor real-user vitals over time. Continuous checks catch both sudden regressions and slow degradations before users feel them.
Read next: rendering and testing tools, or the Resources hub.
Putting it in order?
The performance roadmap shows where each tool fits in the workflow.
Read: Frontend Performance Roadmap →