Skip to content

Alerts

Alerts are generated automatically when a monitored device crosses a threshold, or when Guardian’s predictive engine detects a developing problem. They appear in the Alerts list in real time and refresh automatically.

Severity Meaning
Critical Immediate attention required
Predictive (ML) Guardian’s machine-learning engine flagged a likely upcoming issue before it becomes critical
Warning Worth reviewing soon
Info Informational, non-urgent

Critical and Predictive alerts are highlighted with a pulsing indicator so they stand out in the list.

The Alerts page has three tabs: Active, Resolved, and All. Within a tab you can:

  • Search by alert title or hostname
  • Filter by severity
  • See how many alerts are currently active in the header

Each alert shows its severity, title, source host, an optional message, and the time it was created.

When several alerts fire across different hosts at the same site within a few minutes of each other — the classic pattern for a site-wide event like a switch or power outage — Guardian links the later ones to the earliest as their likely cause. A linked alert shows a ↳ Linked to #N badge pointing at the root alert, so a site-wide event reads as one story instead of a wall of unrelated-looking alerts. This is inventory information only for now — it doesn’t change how alerts are resolved or routed.

Resolving an alert marks it as handled and removes it from the Active view. This requires a Technician or Superadmin role — read-only users can view alerts but not resolve them.

  • Resolve a single alert from its row
  • Select multiple unresolved alerts with the checkboxes and use Resolve selected to clear them in bulk
  • Superadmins can also permanently delete selected alerts

A critical alert automatically opens a tracked incident if one doesn’t already exist for it, and is routed to your configured notification channels. If the alert matches a scheduled or self-healing action, the response is recorded in the remediation log.