Incidents

An incident is one problem on one object, built from the symptoms that keep reporting it. This section covers how incidents are created, how to read one, and how they end.

• The incident lifecycle — the four phases and the outcome, and every state an incident can be in.
• The incident list — status tabs, filters and acting on several incidents at once.
• Reading an incident — what each card on the incident page tells you, and what to look at first.
• What ARGUS detects — the three detectors, and how symptoms become incidents.
• Signals — every finding the agents reported, and why it did or did not open an incident.
• Investigation — what the AI investigation does, what it is given, and how to judge its output.
• Closing an incident — resolution, dismissal, expiry and reopening.
• Tolerations — silencing an accepted condition without ignoring the object.