Overview

RUM view events collect extensive performance telemetry for every pageview. Monitor your application’s pageviews and explore performance telemetry in dashboards and the RUM Explorer.

You can access performance telemetry for your views in:

• Out-of-the-box RUM dashboards, which provide a high-level view of your application’s performance. For example, you can filter on default attributes collected by RUM to surface issues impacting a subset of users in the Performance Overview dashboard. You can also clone this dashboard, customize it to your needs, and use any RUM performance telemetry in the dashboard’s query.
• A performance waterfall, accessible for every RUM view event in the RUM Explorer, which enables you to troubleshoot the performance of a specific page view. It displays how your website assets and resources, long tasks, and frontend errors affect the page-level performance for your end users.

Event timings and core web vitals

Google’s Core Web Vitals are a set of three KPIs designed to monitor a site’s user experience. These KPIs focus on giving you a view of load performance, interactivity, and visual stability. Each KPI comes with guidance on the range of values that translate to good user experience. Datadog recommends monitoring the 75th percentile for these KPIs.

• LCP and FCP are not collected for pages that start hidden (for example, opened in a background tab). For more on when each vital is reported and how to troubleshoot a missing value, see When each Core Web Vital is collected.
• Telemetry collected from your real users’ pageviews may differ from those calculated for pages loaded in a fixed, controlled environment such as a Synthetic browser test. Synthetic Monitoring displays Largest Contentful Paint and Cumulative Layout Shift as lab telemetry, not real telemetry.

Core web vitals target elements

Identifying what element triggered a high Core Web Vitals KPI is the first step in understanding the root cause and being able to improve performance. RUM reports the element that is associated with each Core Web Vital instance:

• For Largest Contentful Paint, RUM reports the CSS Selector of the element corresponding to the largest contentful paint.
• For Interaction to Next Paint, RUM reports the CSS selector of the element associated with the longest interaction to the next paint.
• For First Input Delay, RUM reports the CSS selector of the first element the user interacted with.
• For Cumulative Layout Shift, RUM reports the CSS selector of the most shifted element contributing to the CLS.

When each core web vital is collected

The RUM Browser SDK reports each Core Web Vital under different conditions. If a vital is missing from a view event, the SDK most likely did not capture a value for it.

• Largest Contentful Paint: Reported on the initial view only. Captured when the largest-contentful-paint PerformanceObserver entry fires before the page is hidden and before the user’s first interaction. Not reported if the page starts hidden, if the user interacts before any LCP entry arrives, or after a 10-minute cap. See LCP: Differences between the metric and the API for the upstream definition.
• Interaction to Next Paint: Reported on every view, including SPA route changes. Requires at least one user interaction during the view’s lifetime (see INP on web.dev). If the user never interacts, no INP value is emitted. Requires browser support for the event PerformanceObserver entry type and PerformanceEventTiming.interactionId. If unsupported, no INP value is emitted.
• Cumulative Layout Shift: Reported on every view, including SPA route changes. Requires browser support for the layout-shift PerformanceObserver entry type. The SDK also relies on WeakRef internally; if either is missing, no CLS value is emitted.
• First Contentful Paint: Reported on the initial view only (see FCP on web.dev). Captured when the entry fires before the page is hidden. The SDK applies a 10-minute ceiling per view.

Note: Reporting CLS and INP on every view, including SPA route changes, is intentional. This follows Chrome’s Soft Navigations API experiment, which extends these metrics beyond the initial page load. LCP and FCP use the standard page-load definition and are reported only on the initial view.

Values are captured up to the point of backgrounding; they are not discarded if a page is later hidden. A page that is hidden at view start (for example, opened in a background tab) emits no LCP or FCP.

Troubleshooting missing core web vitals

If a Core Web Vital is missing from a view in the RUM Explorer, check the following:

To check that the SDK is collecting vitals:

1. Open your browser’s developer tools and run window.DD_RUM.getInternalContext() to confirm the SDK loaded.
2. In the Network tab, look for view events being sent to the Datadog intake.
3. In the RUM Explorer, open the view event in question and confirm it has the expected @view.* attributes.

Diagnose Core Web Vitals with subparts

Largest Contentful Paint and Interaction to Next Paint break down into subparts, each isolating a specific phase of the metric. Use subpart data to identify which phase contributes most to a slow LCP or INP.

Largest Contentful Paint subparts

LCP breaks down into four phases. Time to First Byte is collected separately as view.first_byte. The remaining three subparts are collected under view.performance.lcp.sub_parts:

Interaction to Next Paint subparts

INP breaks down into three phases, collected under view.performance.inp.sub_parts:

All performance telemetry

Monitoring single page applications (SPA)

For single page applications (SPAs), the RUM Browser SDK differentiates between initial_load and route_change navigation with the loading_type attribute. If an interaction on your web page leads to a different URL without a full refresh of the page, the RUM SDK starts a new view event with loading_type:route_change. RUM tracks URL changes using the History API.

Datadog provides a unique KPI, loading_time, which calculates the time needed for a page to load. This KPI works for both initial_load and route_change navigation.

How loading time is calculated

To account for modern web applications, loading time watches for network requests and DOM mutations.

• Initial Load: Loading Time is equal to whichever is longer:

  • The difference between navigationStart and loadEventEnd, or
  • The difference between navigationStart and the first time the page has no activity. Read How page activity is calculated for details.

• SPA Route Change: Loading Time is equal to the difference between the URL change and the first time the page has no activity. Read How page activity is calculated for details.

Manually set the loading time

If the automatic loading time calculation does not accurately reflect when your view has finished loading, you can manually set it using setViewLoadingTime. Call this method when your view is fully loaded and displayed to the user. The loading time is computed as the elapsed time since the current view started.

window.DD_RUM.setViewLoadingTime()

Each call replaces any previously set value (last-call-wins). After it is called, the automatic loading time detection is stopped and the manual value is used instead.

After the loading time is sent, it is accessible as @view.loading_time and is visible in the RUM UI.

How page activity is calculated

The RUM Browser SDK tracks the page activity to estimate the time until the interface is stable again. The page is considered to have activity when:

• xhr or fetch requests are in progress.
• The browser emits performance resource timing entries (loading end of JS, CSS, etc.).
• The browser emits DOM mutations.

The page activity is considered to have ended when it hasn’t had any activity for 100ms.

Note: Only activity occurring after the SDK initialization is taken into account.

Caveats:

The criteria of 100ms since last request or DOM mutation might not be an accurate determination of activity in the following scenarios:

• The application collects analytics by sending requests to an API periodically or after every click.
• The application uses “comet” techniques (that is, streaming or long polling), and the request stays on hold for an indefinite time.

To improve the accuracy of activity determination in these cases, specify excludedActivityUrls, a list of resources for the RUM Browser SDK to exclude when computing the page activity:

window.DD_RUM.init({
    ...
    excludedActivityUrls: [
        // Exclude exact URLs
        'https://third-party-analytics-provider.com/endpoint',

        // Exclude any URL ending with /comet
        /\/comet$/,

        // Exclude any URLs for which the function return true
        (url) => url === 'https://third-party-analytics-provider.com/endpoint',
    ]
})

You can also ignore specific DOM mutations by marking an element (or one of its ancestors) with the attribute data-dd-excluded-activity-mutations.
This is useful for elements that constantly update but don’t indicate real UI instability like loading spinners, progress bars, or live clocks.

<div id="loading-spinner" data-dd-excluded-activity-mutations>
  <svg> ... </svg>
</div>

Hash SPA navigation

The RUM SDK automatically monitors frameworks that rely on hash (#) navigation. The SDK watches for HashChangeEvent and issues a new view. Events coming from an HTML anchor tag which do not affect the current view context are ignored.

Create custom performance telemetry

Measure component-level performance with custom vitals

Use the customVital API to measure the performance of your application at the component level. For example, you can measure how long it takes for part of your page to render or for a component to respond to a user interaction. Note: Custom vital names must contain only letters, digits, or the characters - _ . @ $.

Start and stop duration measurements

Start a duration measurement by calling startDurationVital and stop a measurement with stopDurationVital:

window.DD_RUM.startDurationVital("dropdownRendering")
window.DD_RUM.stopDurationVital("dropdownRendering")

Once you call the stopDurationVital method, the custom vital duration is sent to Datadog and can be queried using @vital.name:dropdownRendering. You can also filter by duration, for example with @vital.duration:>10.

Use references and descriptions

Use the reference returned by startDurationVital and specify a description string to differentiate between instances of the same custom vital across multiple pages. For example, to track the duration of dropdownRendering in the login page:

const reference = window.DD_RUM.startDurationVital("dropdownRendering", { description: "login" })
window.DD_RUM.stopDurationVital(reference)

This code groups by @vital.description so you can track the same component’s rendering behavior across different pages.

You can also add context to your custom vital using the context property:

window.DD_RUM.startDurationVital("dropdownRendering", {context: { clientId: "xxx" }})
window.DD_RUM.stopDurationVital("dropdownRendering")

Report a custom vital with addDurationVital

Instead of setting custom vital variables individually, you can report a custom vital in a single operation using addDurationVital:

window.DD_RUM.addDurationVital("dropdownRendering", {startTime: 1707755888000, duration: 10000})

Note: The startTime parameter expects a UNIX timestamp in milliseconds (the number of milliseconds since January 1, 1970).

Track additional performance timings

On top of RUM’s default performance timing, you may measure where your application is spending its time with greater flexibility. The addTiming API provides you with a simple way to add extra performance timing.

For example, you can add a timing when your hero image has appeared:

<html>
  <body>
    <img onload="window.DD_RUM.addTiming('hero_image')" src="/path/to/img.png" />
  </body>
</html>

Or when users first scroll:

document.addEventListener("scroll", function handler() {
    //Remove the event listener so that it only triggers once
    document.removeEventListener("scroll", handler);
    window.DD_RUM.addTiming('first_scroll');
});

Once the timing is sent, the timing is accessible in nanoseconds as @view.custom_timings.<timing_name>, for example: @view.custom_timings.first_scroll. You must create a measure before creating a visualization in the RUM Explorer or in your dashboards.

For single-page applications, the addTiming API issues a timing relative to the start of the current RUM view. For example, if a user lands on your application (initial load), then goes on a different page after 5 seconds (route change) and finally triggers addTiming after 8 seconds, the timing is equal to 8-5 = 3 seconds.

If you are using an asynchronous setup, you can provide your own timing (as a UNIX epoch timestamp) as a second parameter.

For example:

document.addEventListener("scroll", function handler() {
    //Remove the event listener so that it only triggers once
    document.removeEventListener("scroll", handler);

    const timing = Date.now()
    window.DD_RUM.onReady(function() {
      window.DD_RUM.addTiming('first_scroll', timing);
    });
});

Further Reading

お役に立つドキュメント、リンクや記事:

Interactive Lab: Optimizing Frontend Performance with Datadog RUM Browser MonitoringLEARNING CENTER Monitor Core Web Vitals with Datadog RUM and Synthetic MonitoringBLOG Monitoring single-page app interactivity with Core Web Vitals and DatadogBLOG Explore your views within DatadogDOCUMENTATION Apply visualizations on your eventsDOCUMENTATION Learn about RUM DashboardsDOCUMENTATION From performance to impact: Bridging frontend teams through shared contextBLOG Configuring JavaScript caches for better performanceBLOG