← Back to help centre

Incidents

How WebCheck confirms an incident

Understand retries, failure thresholds, incident opening, recovery and why one failed request may not open an incident.

A single failed run does not always open an incident. WebCheck retries eligible HTTP connection failures within one run, then checks the target’s consecutive-failure threshold before opening an incident.

What gets retried?

An HTTP connection failure can be retried every 10 seconds for up to 60 seconds within the same check run. A response such as HTTP 500 or 429 is a completed request, so WebCheck evaluates it without retrying. Safety blocks are not bypassed.

When does an incident open?

The target’s failure threshold sets how many consecutive Failed or Blocked runs are needed. The default is two, and supported targets let you choose a threshold from one to ten. A Passed or Degraded run breaks the failure sequence.

What counts as recovery?

Recovery requires the configured number of consecutive Passed runs. Degraded runs do not recover an open incident. The recovery threshold defaults to one.

Follow the timeline

Compare the first failed run with the last successful run and the later recovery runs. Check the error code, status, timing, redirect chain and available evidence before changing the target.