Skip to main content

Real User Monitoring Metrics

RUM metrics are automatically generated for you from browser traces. They provide insight into your website's overall user experience as well as front-end services, operating systems, geographical locations, and top-loaded page groups and user cohorts categorized by their browsers.

Metrics are collected for user actions representing document loads, which means actual retrieval and execution of web documents in the browser as well as XHR calls and route changes. Measurements include W3C navigation timings, XHR delays, Core Web Vitals KPIs, longtask events (delays) and others.

For ad hoc queries, you can find these metrics in Metrics Explorer by querying for:


Document Load Metrics

The following table has details on the Document Load metrics collected from JavaScript. These are available in each trace in the documentLoad and documentFetch spans as span events in the details panel and also used in the Real User Monitoring app to populate the Website Performance and UI Paint Timings panels.

browser_time_to_dns_resolution_enddomainLookupEnd - span start time (fetch start)
browser_time_to_ssl_endif secureConnectionStart > 0: connectionEnd - span start time (fetch start) else NaN
browser_time_to_tcp_establishedif secureConnectionStart > 0: secureConnectionStart - span start time (fetch start) else connectionEnd - span start time (fetch start)
browser_time_to_request_endresponseStart - span start time (fetch start)
browser_time_to_fb *responseEnd - span start time (fetch start)
browser_time_to_response_enddomInteractive - span start time (fetch start)
browser_time_to_interactivedomComplete - span start time (fetch start)
browser_time_to_processing_endloadEventEnd - span start time (fetch start)

The above metrics, presented in the form of areas on the Website Performance panels on RUM dashboards, can help you understand the sequence of events (pictured below) from user clicks to a fully loaded document.
Diagram showing various navigation metrics from user clicks to a fully loaded document, including DNS resolution, SSL connection, TCP establishment, request end, first byte, response end, interactive time, and processing end

* Time to first byte (browser_time_to_fb): measures the delay between start of the page load and moment when the first byte of the response appears. It helps identify when a web server is too slow to respond to requests. You'll find this metric on the Navigation Timings chart.
Screenshot of the Real User Monitoring interface showing detailed metrics and performance data for web applications

Timing metrics are not calculated if the visibility state of the document was "hidden" at any point during the load.

See W3C navigation timing for details on how an interface for web applications defines its access timing information concerning navigation and other elements.

Rendering Events Metrics

These metrics, which populate in the UI Paint Timings panel on RUM dashboards, explain rendering events inside the user's browser.

browser_time_to_fpfirstPaint - span start time (fetch start)
browser_time_to_fcpfirstContentfulPaint - span start time (fetch start)
browser_time_to_lcplargestContentfulPaint - span start time (fetch start)
  • First Paint: measures the time from page fetch start (span start time) to the moment when the browser renders the first pixels to the screen, rendering anything that is visually different from what was on the screen prior to navigation. It answers the question, "Is it happening?" 
  • First Contentful Paint: measures the time from page fetch start (span start time) to the moment when any part of the page's content is rendered on the screen. For this metric, "content" refers to text, images (including background images), <svg> elements, or non-white <canvas> elements.
  • Largest Contentful Paint: measures the time from page fetch start (span start time) to the moment when the largest image or text block visible within the viewport is rendered.

These are only loosely related to navigation timings and in many cases, some of them may appear long after the page is fully loaded in the browser, which indicates rendering slowdowns.

Core Web Vitals Metrics

Core Web Vitals (CWV) is an initiative by Google that defines web page KPIs. Each CWV represents a distinct facet of the user experience that's measurable in the field and reflects the real-world experience of a critical user-centric outcome.

browser_time_fidFrom CVW First Input Delay API
browser_time_to_lcplargestContentfulPaint - span start time (fetch start)
browser_clsFrom CVW Cumulative Layout Shift API

CWV focuses on three aspects of the user experience: document loading, interactivity, and visual stability. This includes the following metrics (and their respective thresholds):

  • First Input Delay (FID): measures interactivity. To provide a good user experience, pages should have a FID of 100 milliseconds or less.
  • Largest Contentful Paint (LCP): measures loading performance. To provide a good user experience, LCP should occur within 2.5 seconds of when the page first starts loading.
  • Cumulative Layout Shift (CLS): measures visual stability. To provide a good user experience, pages should maintain a CLS of 0.1 or less.
    Visual representation of Core Web Vitals metrics within the Real User Monitoring dashboard, highlighting key performance indicators like FID, LCP, and CLS

The above three CWV KPIs are captured and displayed on Overview dashboards for Document Load action types. Detailed metrics are available in span metadata for every transaction trace.
Example of span metadata for transaction traces displayed in the Real User Monitoring dashboard, providing detailed performance insights

XHR Monitoring Metrics

An XML HTTP Request (XHR) is a form of communication between the browser and the application backend without (re)loading of the page. A typical example is where a page needs to update a ticker of a price automatically or after pressing the “update price” button next to it.

XHR technique is quite often used in single-page apps — apps that load the page once and then provide all interaction and navigation without loading more documents. Pages can generate one or more XHR requests, typically in the form of HTTP POSTs/GETs, related to various user actions on a page. Sumo Logic provides following monitoring coverage for XHR interactions:

Pages can generate one or more XHR requests, typically in the form of HTTP Posts, related to various user actions on a page. What we do with this, is:

  • Measure the following performance timings:

    browser_time_to_first_xhrtime from UI interaction until first HTTP POST appears
    browser_time_to_last_xhrtime from UI interaction until last HTTP POST ends)
    browser_time_to_xhr_processing_endtime from UI interaction until all browser-side processing of all XHR requests is done
    browser_time_in_xhr_callstotal time when the transaction was “busy” with executing XHR communication
  • Measure how many XHR requests have been generated

  • Identify the user action that triggered the XHRs by blending UI interaction (e.g., “click on Pay”) with the page name (e.g., results in following action name "Click on Pay on"

  • Measure any erroneous HTTP response to XHR POST calls and count them as XHR errors

  • Allow to drill-down via EI to specific traces that explain full process of loading and execution of each such transaction
    Interface for monitoring XHR interactions, displaying performance timings and metrics for XMLHttpRequests within a web application

Longtask Delay Metrics

This section describes how to trace and measure Longtask delays, which is when the main browser UI thread becomes locked for extended periods (greater than 50 milliseconds) and blocks other critical tasks (including user input) from being executed.

This impacts the user's experience. They can perceive this as a "frozen browser”, even if the communication with the backend has long been completed. RUM automatically captures such events and:

  • Displays them as individual spans marking how long the browser was frozen
    Graphical representation of longtask delays in the browser UI, showing how these delays impact user experience and are captured in individual spans
  • Aggregates the above data into two metrics:
    • Longtask delay: average duration of longtask span
    • Time in longtasks: total time spent in longtasks per user action/trace
      Aggregated data metrics for longtask delays in the Real User Monitoring dashboard, displaying average duration and total time spent in longtasks per user action or trace.
Privacy Statement
Terms of Use

Copyright © 2024 by Sumo Logic, Inc.