Posted On: Jul 12, 2017
We are pleased to announce that Elastic Load Balancing now supports a new CloudWatch metric called RequestCountPerTarget for Application Load Balancers. The RequestCountPerTarget metric value indicates the average number of requests received by each target in a target group associated with an Application Load Balancer during a specified time period.
This metric is an indicator of load on an individual application backend during a given time window and can be used by Auto Scaling or your own custom scaling system to set up dynamic scaling for your application. This will help you maintain application performance, improve application availability and reduce costs.
This metric is integrated with Auto Scaling through a new type of scaling policy called target tracking scaling policies where you select this metric as a load metric for your application, set the target value, and Auto Scaling adjusts the number of EC2 instances in your Auto Scaling group as needed to maintain that target value. For example, you can configure target tracking scaling policy for your Auto Scaling Group to keep RequestCountPerTarget at 1000 for your webserver fleet. From there, Auto Scaling launches or terminates EC2 instances as required to keep RequestCountPerTarget at 1000.
The RequestCountPerTarget metric is available today for all existing and new Application Load Balancers and can be accessed from the Elastic Load Balancing console or CloudWatch console, SDK and API.