How store health is checked
How the healthy, degraded, unexpected change, failing, and unknown statuses are actually determined for each store.
One store's problem never affects another explains what each health status means in plain language. This page covers how each status is actually determined, rather than just what it looks like once it appears.
What goes into the status
Every health status comes from two things: whether the store matches its declaration exactly, and the results of every check run directly against the live store.
| Status | Requires | | --------------------- | ----------------------------------------------------------------------------------------------------- | --- | | Healthy | Matches exactly, and every check against the live store passed. | | Degraded | Matches exactly, but at least one check came back with a warning. | | Unexpected change | Does not match: the declaration and reality disagree somewhere. | | Failing | At least one check against the live store failed outright, or something that should be running isn't. | | Unknown | Not enough check data came back to decide either way. | |
Every check, accounted for
A store's health also carries a plain summary of how many checks ran and how many passed: something like "4 of 5 checks passing." A "degraded" status is never a mystery; you can always see exactly which check is raising the concern.
Severity and area it affects
Alongside the headline status, you also see which specific area is the cause and how severe the worst issue is: high, medium, low, or none. Nothing is a flat pass/fail with no detail behind it.
Why this matters for your agency
- A status is never a guess. "Degraded" and "failing" come from the same concrete evidence every time, not a fuzzy read of how things seem.
- You know how urgent something is at a glance. Degraded and failing aren't the same situation, and the status tells you which one you're dealing with.
- The detail is right there, not a separate lookup. The area and severity travel with the status instead of requiring you to dig for them.
Where to go next
When unexpected changes trigger an alert
The exact thresholds behind Reqursor's alerts: high severity, unexpected changes lasting over an hour, multiple stores affected, stale checks, and patterns that keep returning.
Fixing unexpected changes from the dashboard or in chat
How you review an AI diagnosis of an unexpected change and approve a fix, through the dashboard or a conversation.