Link Search Menu Expand Document Documentation Menu
Documentation

You're viewing version 2.9 of the OpenSearch documentation. This version is no longer maintained. For the latest version, see the current documentation. For information about OpenSearch version maintenance, see Release Schedule and Maintenance Policy.

Anomaly detection dashboards and visualizations

Introduced 2.9

OpenSearch provides an automated means of detecting harmful outliers and protecting your data when you enable anomaly detection. When applied to metrics, OpenSearch uses algorithms to continuously analyze systems and applications, determine normal baselines, and surface anomalies.

You can connect data visualizations to OpenSearch datasets and then create, run, and view anomaly alarms and results from visualizations in the Dashboard interface. With only a couple of steps, you can bring together traces, metrics, and logs to make your applications and infrastructure fully observable.

Getting started

Before getting started, you must have:

General requirements for anomaly detection visualizations

Anomaly detection visualizations are intended for use with time-series data (real-time or historical). The following visualization types are used frequently for time-series data: line, area, vertical bar, and time-series visual builder (TSVB).

Configuring admin settings

Users can only access, create, or manage alerts for resources for which they have permissions. Access to anomaly detection dashboards and visualizations is controlled by OpenSearch and OpenSearch Dashboards permissions. It is enabled by default and appears as a feature under Dashboards Management > Advanced Settings > Visualization. If the setting is disabled, it does not appear under Dashboards Management. You can disable the setting at the cluster level in the opensearch-dashboards.yml file.

Creating anomaly detectors

To start, first create an anomaly detector:

  1. Select Dashboard from the OpenSearch Dashboards main menu.
  2. From the Dashboards window, select Create and then choose Dashboard.
  3. Select Add an existing, then select the appropriate visualization from the Add panels list. The visualization is added to the dashboard.
  4. From the visualization panel, choose the ellipsis icon (ellipsis icon).
  5. From the Options menu, choose Anomaly Detection > Add anomaly detector.
  6. Select Create new detector.
  7. Input information for Detector details and Model features. Up to five model features are allowed.
  8. To preview the visualization within the flyout, toggle the Show visualization button.
  9. Select Create detector. Once you have created a new detector, the detector is added to the visualization, as shown in the following image.

Interface of adding a detector

Adding anomaly detectors to visualizations

Use a single interface to add, view, and edit anomaly detectors that you want to associate with a visualization. Continuing with the visualization and dashboard in the preceding tutorial, follow these steps to associate an anomaly detector with a visualization:

  1. Select the ellipsis icon (ellipsis icon) from the visualization panel, then select Anomaly Detection.
  2. Select Associate a detector.
  3. From the Select detector to associate dropdown menu, select the detector. Only eligible detectors are listed in the dropdown menu.
  4. View basic information about the detectors. To view comprehensive details, select View detector page to open the Anomaly Detection plugin page.
  5. Select Associate detector. An existing detector is now associated to the visualization, as shown in the following image.

Interface and confirmation message of associating a detector

Refreshing the visualization

Depending on the threshold settings, the visualization refreshes automatically at the specified interval. To manually refresh the visualization, select the Refresh button on the Dashboard page.

Next steps

350 characters left

Have a question? .

Want to contribute? or .