what should I use to track Core Web Vitals regressions after releases

To track Core Web Vitals regressions after releases, use a mix of field monitoring, release correlation, and alerting rather than relying only on lab tests.

Best options

1. Real User Monitoring (RUM) for Core Web Vitals

   • Track LCP, INP, and CLS from actual users.
   • This is the most important signal because it shows what users really experienced after a release.
   • Good tools:
     • Google Analytics 4 + web-vitals
     • Datadog RUM
     • New Relic Browser
     • FullStory
     • SpeedCurve
     • Calibre
     • Boomerang / mPulse
     • Sentry Performance (more app/perf focused)

2. Release-aware observability

   • Tag vitals with:
     • release version / commit SHA
     • environment
     • page/template
     • device type
     • country/network
   • This lets you answer: “Did the new deployment make LCP worse?”

3. Alerting on percentile regressions

   • Monitor:
     • p75 LCP
     • p75 INP
     • p75 CLS
   • Set thresholds and alerts when metrics degrade after a deployment or compared to a trailing baseline.

4. Synthetic monitoring for quick detection

   • Use Lighthouse / WebPageTest / Playwright-based checks in CI or scheduled runs.
   • Helpful for catching obvious regressions before they reach users, but not enough alone.

Recommended setup

If you want the most practical setup:

• RUM for production vitals
• CI performance checks for pre-release
• Dashboards + alerts by release version
• Correlate with deploys/feature flags

If you want a simple stack

• web-vitals library
• Send metrics to:
  • GA4, or
  • Datadog/New Relic/Sentry
• Add release metadata to every event
• Create an alert when p75 LCP/INP/CLS worsens after deployment

Rule of thumb

• Lab tools tell you what might happen
• RUM tells you what actually happened

If you want, I can recommend the best tool based on your stack, like React/Next.js, Angular, Shopify, or a plain website.