RUM Deploys


On this page

Deploys measures every release against the one before it, using real visits to your site. Find it at Site → RUM → Deploys.

RUM Deploys showing UX Rating by deploy, with one regressed deploy.

One shaded area per release, with recent deploys below.

Record deploys with the CLI or API, or connect GitHub.

Read the chart#

Each shaded area is one deploy, measured until the next deploy, for up to 7 days.

  • Dots: one per full day. A deploy shorter than a day shows one dot.
  • Hollow dots: still collecting, or too few page views to compare.
  • Hover: the revision, measurement window, sessions and change.

The chart starts on UX Rating. Switch to LCP, CLS, INP, TTFB, FCP or RTT from the metric menu.

Recent deploys#

  • Deploy: the revision, who shipped it and when. The icon beside the revision opens the commit.
  • Sessions: sessions in the deploy's window.
  • Metric: the value for this deploy, and its change from the deploy before.
  • Collecting: the deploy is still being measured, so its values can change.
  • Low data: too few sessions to compare with the deploy before.
  • Regressed: at least one metric got significantly worse. These deploys are shaded red on the chart, and Regressed above the chart shows only them.

If the page says No Deploys recorded, Calibre hasn't received any yet. See Record deploys. If your latest deploy is older than the date range, the page says when it was and offers a longer range.

Inspect a deploy#

Select a deploy to see:

  • Measured for: how long the deploy was measured.
  • Regressed and Improved: each metric that changed significantly, with its value and change. Select one to see its deploys.
  • Most affected Pages: of the 50 most viewed pages, the ones that changed most, before and after.
A regressed deploy, showing five metrics and their change.

A regressed deploy, with each metric and its change.

For one metric at a time, use the By Deploy chart on each metric page. For lab results, see Synthetic Deploys.

Ask your coding agent#

Ask Calibre MCP “Did yesterday’s deploy make anything slower?” and your agent compares it with the deploy before.