From our ipMonitor guru's Matt and Peter in dev:
When any dependency is ‘down’ ipMonitor will record failures, but those failures will not count towards its alerting (and remedial actions). In this example, while the dependency was down, ipMonitor correctly suppressed alerts while it continued to collect monitoring telemetrics. When the dependency came back up, ipMonitor tested port 80 shortly after and found it still failing. Next, ipMonitor counted that failure towards alerting … and then it kicked off an alerting sequence. Dependencies are about alert suppression.
Here's a link to a tutorial that explains dependencies and alert suppression in more detail: