Setting up AWS Cost Anomaly Detection

We looked at an example use case from our friends at VidyaGames in using Cost Explorer to see anomalous usage in Chapter 3, Managing Inventory. In the example, Jeremy applied grouping and filters to past expenditure data and saw that the Amazon EC2 service and RunInstance API action were contributing significantly to costs. If a bar chart representing daily expenditure shows a sudden spike on a specific day, you can drill down to that day to see what account, service, or API action contributed to the spike. This is a legitimate, albeit manual, approach to detecting unusual expenditure activity.

AWS Cost Anomaly Detection helps to alleviate this manual approach. Cost Anomaly Detection monitors your spend patterns to detect unusual activity using ML. Using this service can help you find anomalies more quickly than doing so manually.

An undesirable effect of anomaly detection is the noise produced from false positives, which tell you that an...