---
title: Configure anomaly alerts
description: Configure anomaly alerts based on metrics outliers on historical patterns in Instance Observer.
locale: en-US
canonical_url: https://www.servicenow.com/docs/r/australia/impact/configure-anomaly-alerts.html
release: australia
topic_type: task
last_updated: "2026-03-12"
reading_time_minutes: 2
breadcrumb: [Anomalies detection in performance metrics, Real-time Instance health monitoring and analysis, Monitor instance performance, Platform Health, Using Impact, Impact]
---

# Configure anomaly alerts

Configure anomaly alerts based on metrics outliers on historical patterns in Instance Observer.

## Before you begin

Role required: admin

## Procedure

1.  Navigate to **Impact** > **Platform Health** > **Monitor** > **Instance Observer**.

2.  From the **Performance** menu, select **Transactions/Response Times**.

    \[Omitted image "io-anomaly-config.png"\] Alt text: Anomaly reporting options configuration section.

3.  Select the reporting options:

    -   Instance
    -   Metrics
    -   Date range
    -   Options:
        -   Self Service Alerts
        -   Diagnostic Events
    -   Add-ons
4.  Select **Get Snapshot**.

5.  To set up an alert, say for a transaction increase, select **Edit Alert**.

    **Note:** Options available on Report and Charts:

    -   Select a data point to view additional actions or remove the entry.
    -   Right-click to isolate a data point.
    -   Select and drag across the chart to drill down.
    -   Download to PNG, SVG, or CSV files or view transactions on the instance.
    -   Change the report type.
    \[Omitted image "io-anomaly-alert-tab.png"\] Alt text: Configure anomaly alerts

6.  Select **Conditions** to choose an IO recommended or an anomaly based condition.

7.  Select the **Snooze time** option, which determines the duration to wait before initiating alerts.

8.  Scroll down to include the supplementary criteria for triggering alerts.

9.  Fine tune the alerts using the listed options.

    -   For example, set a threshold that triggers an alert if the transaction count anomalies persist for more than 10 minutes, or exceeds a certain number of standard deviations from the mean continuously for more than 10 minutes above or below the range, respectively. Job anomalies target top X % of anomalous jobs based on the execution time and track the details of those jobs.
    -   Test and fine tune the alerting by simulating persistence time, choose the window of an anomaly 5, 10, 15 minutes, and so on, or select the top 5,10, or 15 percentage of jobs and verify that alerts are triggered correctly. Adjust the alert threshold, if necessary, to ensure the system provides meaningful and actionable alerts.
    -   Choose the appropriate method for sending alerts based on your requirements. This can include email notifications, SMS messages or integration.
    -   Continuously monitor the alerts generated by the system. Regularly review the anomalies detected to understand the underlying causes and take appropriate actions, such as investigating potential issues or performing further analysis.
10. Select **Notification** to choose a notification, similar to the standard alerts procedure.

11. Select **Save**.

    From now on, you will receive an alert whenever anomalies related to transaction increases are detected.


**Parent Topic:**[Anomalies detection in performance metrics](https://raw.githubusercontent.com/ServiceNow/ServiceNowDocs/australia/markdown/australia/impact/anomalies-detection.md)