Skip to content

Recent Failures

The Recent failures page, in the left menu, lists every check of your team that failed at least once in the selected period, the ones still failing first. It answers "what broke lately?" in one place, on every plan, without opening each check.

For each check you see:

  • the Status: Failing when its latest run failed, Recovered when a later run passed;
  • the number of Failed runs in the period;
  • the First failure and Last failure times.

A check that failed from some of its locations but not from all of them is marked Failed from some locations only, so a single-region blip reads differently from an outage.

Use the Period selector to look at the last 24 hours, 7 days or 30 days. The list covers the first 50 checks of the team; a team with more checks sees a note saying so. The page opens on 7 days.

A run counts as failed when the check failed from at least one of its locations. SLA reports and dashboards use the same definition, so the numbers on the three pages match. A check with no failed run in the period is not listed, and when nothing failed the page says "No check failed in this period."

Failed runs shows the failed runs out of all runs in the period, for example 3 / 288. Click a check's name to open its results page and see each run.

Checks that are still failing come first, then the others by their most recent failure.

  • Failing without the "Failed from some locations only" note: the latest run failed, and at least one run failed from every location at once. Look at these first.
  • Failed from some locations only: every failed run still passed somewhere, which points at a regional problem, such as one CDN edge or one network route, rather than at the service itself.
  • Recovered with one or two failed runs is often a short blip. If blips like that raise alerts you do not need, see retries and alert thresholds.