Posted On: Nov 19, 2018

Starting today, you can use AWS Cost Explorer’s improved forecasting engine to access more accurate spend forecasts that are tailored to your needs, regardless of your environment’s complexity. AWS Cost Explorer’s improved forecasting engine segments your historical data based on distinct charge types (e.g., on-demand usage, reserved instance usage, and more) and uses a combination of machine learning and rules-based models to predict spend across all of those charge types individually. By using this multiple time series-based approach, Cost Explorer ensures a higher degree of accuracy, irrespective of the AWS services being used.

To get started using your improved cost forecasts, navigate to any AWS Cost Explorer report and select a date in the future. Depending on your forecasting needs, you can access forecasts at a high level (e.g., forecasted monthly costs for the rest of the year) or at a more granular level of detail (e.g., forecasted costs for i3.metal instances with the “project:Megatron” tag applied). Forecasting capabilities have also been enhanced to support twelve month forecasts (previously forecasts were limited to three months) for multiple cost metrics, including unblended and amortized costs. Forecasted data is also accessible programmatically via the AWS Cost Explorer API.

To get started with the new forecasting functionality, visit Cost Explorer here. To learn more about AWS Cost Explorer, you can access the AWS Cost Explorer webpage or refer to the Analyzing your costs with Cost Explorer user guide.