How incidents work
When one opens, when it closes, what the times mean, and how to raise one yourself.
An incident is the record of something being down: when it started, when it recovered, and how long that was.
Automatic incidents
One opens when a monitor's consecutive failures reach its failure threshold, and closes on the first success after that.
Started at is the moment the failure was confirmed — when the threshold was reached — not the first failed check. With a threshold of 3 on a 1-minute interval, the first failure was about two minutes before the start time.
That is deliberate: until the threshold is met we do not yet believe it is an outage, and an incident that back-dated itself to a failure we had decided to ignore would make the duration disagree with the alert you received.
Manual incidents
Raise one yourself when something is wrong that no monitor can see — degraded performance, a partial failure, a problem a customer reported before your checks noticed. They are also what appears on your status page, which is usually the real reason to create one.
A manual incident moves through four statuses:
| Status | What it means |
|---|---|
| Investigating | Something is wrong and we are finding out what. |
| Identified | We know the cause and are working on it. |
| Monitoring | A fix is in; watching to confirm. |
| Resolved | Done. |
Each change can carry an update — a line of text that appears on the status page with a timestamp. Those updates are what turn a status page from a red light into something your customers will actually trust.
What the list shows
Open incidents first, then resolved, newest first. Each shows its monitor, when it started, how long it lasted and its status.
Incidents and uptime
An incident's duration is not the same as the gap in your uptime figure. Uptime counts checks: failed checks over total checks. The incident is a span of time.
They disagree when a monitor is flapping — up and down and up again produces several short incidents and a modest uptime dent, because many checks in between passed.
Maintenance
Checks still run and results are still recorded during a maintenance window, but nothing alerts and no incident opens. Those results are excluded from uptime.
Last updated 4 October 2026
Still stuck?
If this did not answer your question, tell us and we will fix the page as well as answer you.