Measured endpoint health

Success rate and typical response time are computed from our own request logs, grouped by catalog category, over the last hour, the last 24 hours and the last 7 days. The short windows surface a live incident; the 7-day window is where lower-volume groups reach a reportable sample. Figures cover customer traffic only: our own testing and audit accounts, withdrawn services, and rate-limited or insufficient-credit responses are all excluded — they describe our staff or an account, not an endpoint. Below 20 calls in a window we publish no percentage, because at that size it is noise rather than a measurement.

This page is served by the same infrastructure it reports on, so it is not an independent oracle: if the API is unreachable, this page may be too. Your own per-request status codes and latency are always in your dashboard request logs.