Every check on this page runs when you load it, against the live API — nothing here is cached or scheduled. If a check cannot run, the page says so rather than reporting the last good answer.
Running live checks…
Every check on this page runs when you load it, against the live API — nothing here is cached or scheduled. If a check cannot run, the page says so rather than reporting the last good answer.
Checked 29 Oct 2026, 17:22 UTC — when you loaded this page, not on a schedule. Build e382513.
Server processing time across 48 live requests in the last 24 hours — handler to response, excluding network transit. A repeat lookup answers from cache; a first-time one runs live MX, SPF, DMARC and RDAP queries against other people's servers, so the two are reported separately below rather than averaged into a number that is true of neither.
Total processing · 48 of 48 requests measured
Repeat lookups · served from cache · 38 of 48
First-time lookups · live registry and DNS queries · 10 of 48
Bars use a logarithmic scale from 10 ms to 1.0 s, so the fast requests keep their variation instead of flattening against the slow ones. 4 of 48 ran longer than 1.0 s and are drawn at full height. No third-party wait was recorded across these requests, so the figure above is the whole request.
If either of these is down, scoring is down.
2 sources, read from the primary publisher in every case — never resold from an aggregator.
4 sources, read from the primary publisher in every case — never resold from an aggregator.
10 sources, read from the primary publisher in every case — never resold from an aggregator.
1 source, read from the primary publisher in every case — never resold from an aggregator.
1 source, read from the primary publisher in every case — never resold from an aggregator.
2 sources, read from the primary publisher in every case — never resold from an aggregator.
3 sources, read from the primary publisher in every case — never resold from an aggregator.
Liveness says a feed answered; these say how much it is worth. Recomputed weekly — as of 25 Oct 2026.
A probe runs against the API and database every five minutes and records the result — including the failures, because an uptime table that only stores successes reports 100% forever. The figure above is that record over the window and sample size shown, rounded down: 99.994% displays as 99.99%, never 100%. We show it from the very first check rather than waiting for a flattering sample, which is why the number of checks is printed next to it — that is what lets you judge how much weight it deserves.
Every check on this page runs live, on this request, with the time it took. If a feed is dark it is named above.
The raw JSON is at /api/status — point your own monitor at it. It returns 503 when core systems are down, so you can alert on the status code without parsing anything.