UXMeasurementList
The UX measurements of a window of hours, by metric and then by most samples.
object
The start of the first hour of the window, in UTC.
The samples of one metric and name over the window.
object
What a UX measurement of the web client measures, and so what its name names. The values are
milliseconds, except call_packet_loss, a percentage from 0 to 100.
lcp,inp,long_frame: the largest contentful paint, the interaction to next paint, and a long animation frame of a route, named by the route’s ID (a pattern such as/_app/statuses/$id, never the URL).transition: a navigation, named by its interaction kind (push,pop,modal,replace,none).press: the time from a press to its response, named by the action ID (post.reply).call_rtt,call_jitter,call_packet_loss: the quality of a voice call, named by the call transport (p2p,livekit,sora).
The route ID, interaction kind, action ID, or call transport; other for the names past
the first 200 of the metric that a process of the server kept.
The upper bound of the bucket that holds the 75th percentile: at least three quarters of the samples are at most this value. Absent when the 75th percentile is in the last bucket.
The samples by bucket, from the lowest bound up.
The samples whose value is above the bound of the bucket before and at most le. The last
bucket has no le: it holds the samples above every bound.
object
The upper bound of the bucket, in the unit of the metric.
Example generated
{ "since": "2026-04-15T12:00:00Z", "items": [ { "metric": "lcp", "name": "example", "samples": 1, "p75": 1, "buckets": [ { "le": 1, "samples": 1 } ] } ]}For LLMs and coding agents:llms.txt,llms-full.txt, llms-small.txt