# Anomalies

## Setting Up Anomaly Detection

Automatically monitor your metrics and get notified when patterns deviate from normal behavior.

{% hint style="info" %}
Anomaly detection requires a minimum of 30 days of time series data. Once enabled, it will run across all areas and opportunities.
{% endhint %}

***

### Detection Types

Birdie offers two types of anomaly detection. Choose the one that fits the pattern you want to track.

{% columns %}
{% column %}
**Spike Detection**

Identifies unusual spikes in a metric compared to a baseline period. Best for catching sudden, sharp increases in volume or frequency.
{% endcolumn %}

{% column %}
**New Pattern Detection**

Identifies when a metric's trend direction changes, for example from upward to downward movement. Best for detecting sustained shifts in behavior over time.
{% endcolumn %}
{% endcolumns %}

### Creating an Anomaly Rule

{% stepper %}
{% step %}

#### Open Settings

Go to Settings and select Anomalies from the left navigation under the Notifications section.
{% endstep %}

{% step %}

#### Add a new anomaly

Click + Add new anomaly in the top right corner.
{% endstep %}

{% step %}

#### Name the anomaly

Enter a name in the Anomaly name field. Choose a name that clearly identifies the metric and behavior being monitored.
{% endstep %}

{% step %}

#### Enable anomaly detection

You can disable the anomaly detection if you want to just set it up. Later, after it's saved, you can enable it back on the Anomalies settings page.
{% endstep %}

{% step %}

#### Select the detection type

Under Detection type, choose one of the options:

**Spike Detection** monitors for unusual spikes in a metric compared to a baseline period. Best for catching sudden increases in volume.

**New Pattern Detection** monitors for changes in trend direction. Best for identifying sustained shifts in behavior.

<figure><img src="/files/lwwDyMIAcixrPfvywnQw" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}

#### Choose the source to monitor

Under Source to monitor, select the source you want to track for anomalies. By default this is set to All.
{% endstep %}

{% step %}

#### Set the sensitivity

Under Sensitivity, choose when the detection should trigger. The default value is Notable changes, which balances noise reduction with timely detection.
{% endstep %}

{% step %}

#### Configure the delivery channel

Under Delivery Channel, select where you want to receive alerts. You can select one or more users or channels.
{% endstep %}

{% step %}

#### Select the time zone

Under Time, confirm your preferred time zone for receiving alerts. **Alerts are sent at 8:00 AM in the selected time zone.** The default is GMT 3:00 Brasilia Time.
{% endstep %}

{% step %}

#### Save the anomaly detection

Review your configuration and save the rule. Anomaly detection will begin running according to your settings.
{% endstep %}
{% endstepper %}

### Managing Anomaly Rules

Once created, anomaly rules appear in the Anomalies settings page.

**Find by name:** Use the search bar to locate a rule quickly.

**Filter by type:** Use the All types dropdown to filter by detection type.

{% hint style="warning" %}
If no anomaly rules appear, none have been configured yet. Click + Add new anomaly to create your first rule.
{% endhint %}


---

# Agent Instructions: Querying This Documentation

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

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://ask.birdie.ai/admin-and-settings/notifications/anomalies.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
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.