Anomalies

Anomalies enables real-time analysis of a user's cloud costs through user-defined rules and AI algorithms. This allows users to investigate the anticipated reasons behind the occurrence of abnormal costs and easily find solutions to address them.

Checking overall anomaly status

In the Overview tab of Anomalies, you can monitor abnormal cost-related alerts by Cloud Service Provider (CSP). Additionally, you can compare cost trends over the last 7 days and the trends from 7 to 14 days before the current date. By hovering over the AI-based anomaly cost detection trend chart, a tooltip will appear on the screen. Clicking on 'View Details' in the tooltip allows you to access detailed information about AI-based anomaly cost detection over the last 7 days.

Note

"Last 7 days" shows data from the past 7 days, counting from the current date. "Previous 7 days" refers to the data from 7 days before to 14 days before the current date, encompassing a period of 7 days.

Setting alerts for anomalies

In the Alerts tab, you can view detailed information of the configured alert rules, as well as modify or delete existing rules.

1. Click on the 'Add' button in the top right corner to navigate to the rule configuration screen.

2. Select values by refering to the table below.

Item	Category	Description
Analysis Type	Common	Select whether to analyze cost anomalies based on user-defined rules or through AI-based analysis.
Analyze by	Manual	Select whether to analyze cost anomalies by account, by product, or by region.
Period	Manual	* Last 3 days vs. Previous 3 days: Compare and analyze the costs from the last 1-3 days with the costs from the last 4-6 days. * Last 7 days vs. Previous 7 days: Compare and analyze the costs from the last 1-7 days with the costs from the last 8-14 days. * Month-to-date vs. Same period last month: Compare and analyze the costs from the beginning of this month until now with the costs from the same period last month. * Last collected date vs. 7-day average: Compare and analyze the costs from the last recorded date with the average cost over the past 7 days.
Threshold	Common	Enter the amount or percentage for triggering alerts.
Status	Common	This is the option used to temporarily turn off anomaly alerts. It is set to Active by default, and can be changed to Inactive later to switch off the alerts.
Frequency	Manual	Specify how often you would like to receive an anomaly alert.
Options	Manual	Select whether to receive alerts for newly detected anomalies only or for all anomalies that have not yet been resolved.
Recipients	Manual	Enter up to 10 email addresses for alert recipients.
Sensitivity	AI	Set how sensitively anomalies are detected. A higher sensitivity means anomalies are detected with stricter criteria, which would trigger more alerts as a result.

Viewing the history of anomaly detection

Click on the down-pointing angle bracket of an alert rule in the table to inspect specific information about abnormal costs detected by that rule. For manual alerts, you can examine the numerical and graphical representation of how costs have changed and view the costs incurred on each date. Additionally, you can check the results of alert email transmissions. For AI-based alert rules, you can explore information about the AI models that detected abnormal costs, cost changes, cost trends, detailed information about abnormal cost detection, and the results of alert email transmissions.

---