RUM and Synthetic Monitoring

1. Two complementary approaches #

Real and simulated.

Comparison #

2. RUM #

The browser.

Installation #

import { datadogRum } from '@datadog/browser-rum';

datadogRum.init({
  applicationId: '<APPLICATION_ID>', clientToken: '<CLIENT_TOKEN>', site: 'datadoghq.eu',
  service: 'shop-front', env: 'prod', version: '2.3.0',
  sessionSampleRate: 100, sessionReplaySampleRate: 20,
  trackUserInteractions: true, trackResources: true, trackLongTasks: true,
  allowedTracingUrls: ['https://api.example.com'],     // relie les requêtes aux traces backend
  defaultPrivacyLevel: 'mask-user-input',
});

• Core Web Vitals (LCP, INP, CLS) per page, browser and country.
• Session Replay to replay a session, with input masking (defaultPrivacyLevel).
• allowedTracingUrls adds tracing headers to API calls: in one click, you go from a slow action to the backend trace.

3. Synthetics #

Testing continuously.

Test types #

• Managed locations (cloud regions) or private locations (a container in your network) to test internal applications.
• Running tests in CI (datadog-ci synthetics run-tests) to block a deployment that breaks a journey.
• Tests and locations managed in Terraform (datadog_synthetics_test).

Good practice: Alert when several locations fail (for example 2 out of 3) rather than just one: a local network issue should not wake up on-call.