Tutorial - Autoscaling using requests per second

Setting up microscaling based on requests per second

IMPORTANT: Tutorials are intended to give you hands-on experience working with a limited set of DC/OS features with no implied or explicit warranty of any kind. None of the information provided—including sample scripts, commands, or applications—is officially supported by Mesosphere. You should not use this information in a production environment without independent testing and validation.

You can use the marathon-lb-autoscale application to implement request rate-based autoscaling with Marathon. The marathon-lb-autoscale application works with any application that uses TCP traffic and can be routed through HAProxy.

marathon-lb-autoscale collects data from all HAProxy instances to determine the current RPS (requests per second) for your apps. The autoscale controller then attempts to maintain a defined target number of requests per second per service instance. marathon-lb-autoscale makes API calls to Marathon to scale the app.

For more information, see the Marathon-LB Reference.