DoiT Anomaly Detection can now identify unplanned cost spikes at both the service and SKU levels simultaneously, without duplicating the alerts.

Previously, our anomaly detection and alert granularity was at the desired SKU-level to speed up your time-to-resolution when there was an issue. However, when a new SKU was introduced, it took several days to recognize any anomalies due to insufficient data for the machine learning model to analyze effectively. With the recent improvements, we can now detect unplanned cost spikes at the service level, even if they go unnoticed at the SKU level due to data limitations.

Anomalies detected at the service level are automatically flagged and alerted, mirroring the existing functionality for SKU-level cost anomalies. The only visual distinction is that the SKU field will display “ALL” on the list page, view page, and alert notifications.

Anomaly list screen

Anomaly view screen

To learn more about anomaly detection, check out our documentation in the Help Center.