> ## Documentation Index
> Fetch the complete documentation index at: https://docs.usestealth.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# Incidents

> Every open and seen incident of the organization, and the ones resolved in the last 30 days.

An incident is a drop that Stealth detected and is following: it opens when the drop appears, can get worse, and is resolved when the numbers return to normal. **Incidents** lists them all in one place. Click the type of an incident to open its [Incident page](https://docs.usestealth.dev/en/screens/incident.md), with the key numbers, the possible causes and the team verdict. To learn how incidents are opened and resolved, see [Incident lifecycle](https://docs.usestealth.dev/en/detection/incident-lifecycle.md).

## What you see

The page opens on **Open** incidents. Use the filter at the top to show **All**, **Seen** or **Resolved** incidents instead. Resolved incidents stay on this page for 30 days.

| Column | What it shows |
| - | - |
| **Type** | The kind of drop and its severity (**Critical**, **Attention** or **Light**). |
| **Client** | The client of the incident. |
| **Network / block** | The network and the ad unit, or the whole client when the incident covers the client. |
| **Started** | The hour the drop started. |
| **Estimated loss** | Expected revenue minus real revenue during the incident. It is an estimate, not a measured value. |
| **Status** | **Open**, **Seen** or **Resolved**. |

### Types

| Type | Meaning |
| - | - |
| **Ad unit stopped** | An ad unit stopped receiving requests while the rest of its client kept working. |
| **No ads served** | An ad unit still receives requests but almost none is filled. |
| **Client lost traffic** | A client's requests fell to less than half of normal, beyond the market. |
| **Client fill rate dropped** | A client's fill rate fell below 60% of normal while its traffic held. |
| **Revenue per request dropped** | A client earns much less per request than in the same weekdays of previous weeks. |
| **Revenue trending down** | A client's revenue of the last 7 days is well below the 14 days before, with no equivalent drop in traffic. |

The exact rules are in [Alert kinds](https://docs.usestealth.dev/en/detection/alert-kinds.md). Older incidents may show types from earlier versions of the detection, such as **Revenue zeroed** or **Network-wide drop**.

### Mark as seen

Any member can choose **Mark as seen** on an open incident. The status changes to **Seen** for the whole team, so others know someone is looking at it. Detection goes on as before: the incident can still get worse or be resolved, and alerts keep following the rules.

## FAQ

> [accordion] Where does the link in an alert email or webhook take me?
>
> To the [Incident page](https://docs.usestealth.dev/en/screens/incident.md). An alert email that covers several incidents of one client opens the client page instead, and one that mixes clients opens the Panel. If the incident belongs to another organization you are a member of, the dashboard switches to it.

> [accordion] Can I close an incident by hand?
>
> No. An incident is resolved only when the numbers return to normal. To stop alerts about something you already know, create a [mute exception](https://docs.usestealth.dev/en/alerts/overview.md).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.