Anomalies - Optimaze

OPTIMAZE watches spend for changes that break from its recent pattern, a sudden jump from one day to the next, a new service appearing on the bill, a slow drift that no single day would reveal. Each one becomes an anomaly, so you see the change before it reaches the invoice.

The Anomalies page

Under Explore, select Anomalies. Each anomaly shows:

Field	What it tells you
Title	What changed: a Service, an Account, a Region, or Total Cost for a whole integration
Integration	Where the spend comes from, such as AWS or Cursor
Increase or Decrease	The direction of the change
Description	What was detected, for example a new service or a change in the seven-day average
Change	The net dollar change, the percentage change and the value it changed from
Last Detected	When the anomaly was last seen
Severity	How significant the change is, such as Warning or Info
Detection Window	The comparison that found it (below)

Anomalies page in card view, with account, service and total cost anomalies and the filters above

Detection windows

Window	Compares
Day over Day	One day with the day before
Week over Week	One week with the week before
Month over Month	One month with the month before
Quarter over Quarter	One quarter with the quarter before
Gradual Drift	A rolling average with an earlier period, to catch slow changes

Filter and sort

Service: find anomalies for a service by name.
Integration: limit to one integration. Where you have several integrations from one provider, totals are named by integration.
Detection Window: show one window, or All.
Trend: show all anomalies, only increases, or only decreases. Each option shows how many anomalies it holds.

Sort by net change, and switch between card and table views.

Get notified

Connect Slack under Data > Integrations to send anomaly notifications to a Slack channel.

Investigate an anomaly

Open Analytics filtered to the same integration and service to see the spend behind it, or select Atlas AI and ask why the cost changed.