Home

Anomaly Detector API Overview

Anomaly detection involves identifying unexpected events, items, or observations in data sets that are suspicious because they differ significantly from standard behaviors or patterns. The anomaly detector API gathers metrics from and returns anomalous data.

Consider the following example where you run the Databricks jobs for a specific time, and the costs in each Databricks job usually run between 100 to 500 dollars range. Then, suddenly the cost rises to 1200 dollars or sometimes goes down to 30 dollars. This behavior is an unexpected event and is called an anomaly. By using the anomaly detection API, you can get an alert of such unexpected events in your data sets.

anomaly-chart.png