Anomaly Identification
Anomaly Identification policies are similar to budgets in that they will track and alert on cost and resource count violations, using the same filtering approach as budgets for scoping. Anomaly polices, however, are evaluated over rolling historical period (last x days) against a percentage threshold.
In this image we can see the description of a spend anomaly policy and a resource count policy:
The spend (opex) policy example above will be violated if the spend for a new day exceeds the average daily spend over the last 7 days by thirty percent.
The main Anomaly Identification screen lists your policies, and each row includes a mini chart showing the spend or resource count (in scope) over the last x days from today, where x is the rolling period defined in the policy. Red indicates a violation:
Clicking on a policy opens a details page, allowing you to rename or delete the policy. Also, on the page you can see violation history and a chart at the top-right where you can start to explore policy violation root causes.
Note that once a policy is configured, you can’t edit the threshold or filters, as this would break the validity of the policy violation history.
Email notifications will be received for policy violations.
The Executive Dashboard includes a Policy Violations tile at the lower-left that lists policy violations and includes drill-through links.