AutoScalr provides a machine learning powered engine as a service to make the decisions about scaling your AWS cloud application in the most cost-effective manner. You run a simple CloudFormation template that sets up a secure cross-account role access per AWS Best Practices, specify your target auto-scaling group, the instance types, and the maximum capacity you want your application to use spot instances for through either CloudFormation, Terraform, or Web UI, and we do the rest.
Once enabled, AutoScalr continuously monitors your application’s load, CPU levels, spot market prices, and your diversity across different spot markets to make sure your specifications are always met.
When necessary, AutoScalr will issue scaling commands on your behalf, selecting the most appropriate spot market (instance type and availability zone) and/or on-demand or reserved instances, based on multiple factors. All automatically.
The end result is you save money. A lot of money. Typically between 50-75% of your current cost of running a load balanced autoscaled application on AWS, while maintaining and often increasing the performance and availability of your application because we can frequently run more instances at a lower price point.