Overview

What Is Cost Anomaly Detection?

Cost Anomaly Detection uses machine learning to analyze your historical pay-per-use and yearly/monthly expenditures, establish a specific expenditure model for you, and identify root causes for cost surprises based on forecasted amounts. With simple steps, Cost Anomaly Detection helps you quickly take action based on detected cost anomalies to maintain your planned expenditures.

You can create monitors for all services, for just linked accounts, or based on cost tags. Only one monitor type is recommended for an account. Otherwise, duplicate anomalies may be recorded.

• All services: This type of monitor tracks the expenditure anomalies for all your services. It is recommended if you do not need to group costs within your enterprise. Only one monitor of this type can be created under an account.
• Linked accounts: This type of monitor tracks the pay-per-use expenditure anomalies for an individual linked account. It can be useful if you are using a master account and want to group costs by linked accounts. The master account can create only one monitor of this type for each linked account.
• Cost tags: This type of monitor tracks the expenditure anomalies for an individual cost tag key-value pair. It is recommended if you want to group costs by cost tags. Only one monitor of this type can be created for each cost tag value.
• Enterprise projects: This type of monitor tracks pay-per-use and yearly/monthly expenditure anomalies for the specified enterprise project. It is recommended if you want to group costs by enterprise project.

Detection Rules

Cost Anomaly Detection helps you monitor the actual payments of both pay-per-use and yearly/monthly resources.

• Pay-per-use expenditures: AI algorithms are used to intelligently identify unexpected expenditure spikes based on machine learning.
• Yearly/Monthly expenditures: A cost anomaly is identified if the month-to-date (MTD) expenditures have increased by a certain percent over the previous billing cycle.

  For example, if you set the percent to 20%, then a cost anomaly will be identified if the actual growth rate is higher than 20%. Actual growth rate = (Actual cost for the current month – Cost for the previous month)/Cost for the previous month