Cost anomalies grouped by incident, with alerts for the big ones

See every detection from an incident together, and hear about the ones that grow.

A cost spike rarely shows up as a single anomaly. One incident can trigger detections across several SKUs or projects at once, and the first one detected isn't always the one that matters most.

Cost Anomalies now brings every related detection into one group, and alerts you when one of them outgrows what you already know about.

• Alerts when a related anomaly grows: A related anomaly gets its own alert once its excess cost reaches 10x the amount already alerted within that group, or exceeds it by $1,000. The next one only fires if costs climb much further.
• One row per incident: Related anomalies appear as a single collapsible row in the list. Expand it to see every detection in the group and open any of them.
• Linked Anomalies on every detail page: A new table lists the other detections in the same group, so you can cross-reference without leaving the page.
• Service vs. SKU level: A new "Level" column shows whether each detection is at the service or SKU level, so you can triage faster.
• Group-aware filters, exports and API: Filters and sort order keep related anomalies together. CSV exports and the external API include every group member.

To see it, open Cost Anomalies and expand any grouped row.