> For the complete documentation index, see [llms.txt](https://docs.itoc360.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.itoc360.com/on-call-management-and-escalations/escalations/escalation-grouping.md).

# Escalation Grouping

When a system starts having problems, it rarely fires a single alert. A database going down might trigger dozens of alerts within seconds, one for each dependent service, check, or threshold.

Without grouping, each of those alerts would kick off its own escalation chain and wake up your on-call engineer multiple times for what is fundamentally the same incident.

Escalation Grouping solves this by collecting the alerts a policy receives into a single incident before notifications start.

Grouping works per policy, not by alert content. Every alert that matches the policy's rule joins the same open group, no matter what the alert is about. If you want different kinds of alerts grouped separately, create separate policies, each with its own rule. See [Rule Engine](https://docs.itoc360.com/on-call-management-and-escalations/escalations.md#rule-engine).

### Configuring Grouping

To set up grouping for an escalation policy:

1. Navigate to Management > **Escalation policies** and open the policy you want to configure.
2. Switch to the **Grouping** tab.
3. Click **Edit**.

In the **Edit grouping** dialog, choose a mode and set its limits, then click **Save changes**. The dialog shows a sample timeline of how alerts would be grouped with your settings. Changes apply to new alerts; a group that is already collecting alerts keeps its timer.

### Grouping Modes

There are three modes to choose from, depending on how your alert patterns behave.

#### Off

Every matching alert opens its own incident and notifies right away. This works well for low-volume, isolated incidents where every alert is expected to be distinct.

#### Fixed window

The timer starts when the first alert arrives and runs for the **Window** you set. Alerts that arrive during the window join the same incident. New alerts don't extend the window. When it runs out, the group closes and notifications start.

Use Fixed window when you know your alert bursts are short-lived and predictable. It gives you a hard boundary on how long alerts are held before someone gets notified.

#### Sliding window

Each new alert restarts the timer. The group closes once no new alert has arrived for the full **Quiet period**, or when it reaches the **Max duration**, whichever comes first.

Use Sliding window when your incidents tend to produce waves of alerts with short gaps in between. It keeps everything under one incident until the situation actually settles down.

**Fixed window** and **Sliding window** are plan features. If your plan doesn't include one, it appears locked in the dialog with "Not in your plan". If a policy already uses a mode your plan no longer includes, it keeps working, but once you switch to another mode you can't switch back.

### Window and Timing

#### Window and Quiet period

Set in seconds, from 10 to 3600. The default is 60.

* **Window** (Fixed window): how long the group stays open after the first alert.
* **Quiet period** (Sliding window): how long to wait without a new alert before the group closes. Each new alert restarts this wait.

Adjust this based on how long your typical alert bursts last.

#### Max duration

Sliding window only. The longest a group can stay open, counted from the first alert, in seconds (up to 14400). It can't be shorter than the Quiet period. Without it, a steady stream of alerts could keep the window open forever. The default is three times the Quiet period.

#### Max alerts

The most alerts one group can hold, from 2 to 1000 (default 100). When the group reaches it, it closes right away and notifications start without waiting for the timer. The next alert opens a new incident.

This acts as a safety valve for unusually large incidents where waiting for the timer would delay the response too long.

### Choosing the Right Mode

A few practical guidelines:

* If you receive highly critical alerts from your monitoring tool and don't want to wait for any grouping window before getting notified, **Off** is the right choice.
* If you see short bursts of alerts during incidents, **Fixed window** is usually the right call. It's simpler and more predictable.
* If your incidents roll in waves, with alerts firing, pausing, then firing again, **Sliding window** keeps them grouped until things actually calm down.

When in doubt, start with **Fixed window** (if your plan includes it) and adjust the Window based on what you see in practice.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.itoc360.com/on-call-management-and-escalations/escalations/escalation-grouping.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
