Introducing AWS Cost Anomaly Detection (Preview)

Starting today, you can receive anomaly detection alert notifications with root cause analysis, so you can proactively take actions and minimize unintentional spend. AWS Cost Anomaly Detection is backed by a machine learning model that is able to detect various types of anomalies (whether a one-time cost surge, or gradual cost increases) with minimal user intervention. The model learns your historical cost and usage, as well as accounts for unique, organic growth and seasonal trends. AWS Cost Anomaly Detection goes further by providing root cause analysis when possible, saving you from spending time to research these anomalies. As part of the [AWS Cost Management suite](/aws-cost-management/), AWS Cost Anomaly Detection is integrated with AWS Cost Explorer so you can further visualize and analyze your cost and usage as needed. Additionally, you can subscribe to email or SNS notifications that will alert you based on your pre-defined alert preferences. To get started, click on Anomaly Detection listed in the AWS Cost Management sidebar and opt-in to this feature. You can set up granular Anomaly Detection by creating Monitor Types, such as AWS Service, Account, Cost Allocation Tag, or Cost Categories. After you configure the alerting preferences, Anomaly Detection may take up to 24 hours to build tailored models and begin evaluating cost and usage for anomalies. To learn more about AWS Cost Anomaly Detection, you can reference the [Cost Anomaly Detection page](/aws-cost-management/aws-cost-anomaly-detection/), [Cost Anomaly Detection user guide](https://docs.aws.amazon.com/awsaccountbilling/latest/aboutv2/manage-ad.html), and the [getting started blog post](https://aws.amazon.com/blogs/aws-cost-management/preview-anomaly-detection-and-alerting-now-available-in-aws-cost-management/). To access capabilities programmatically, you can visit the [AWS Cost Explorer API](https://docs.aws.amazon.com/aws-cost-management/latest/APIReference/API%5FOperations%5FAWS%5FCost%5FExplorer%5FService.html) documentation.