Auto Scaling cloud resources is nothing new in AWS. However, Amazon recently announced a new Target Tracking policy that gives customers more granular control over how their application scales. Target Tracking policies allow an administrator to target a specific metric that will drive how and when the EC2 resources will scale. This approach is an alternative to traditional methods which rely upon ranges and thresholds.
Auto Scaling is one way organizations can maximize the value of the AWS cloud. Jeff Barr, chief evangelist at AWS, explains why it is important:
Our new Auto Scaling allows you to build systems that respond to changes in demand by launching or terminating Amazon resources based upon conditions you define. This dynamic scaling helps you improves application availability and reduce costs.
Image Source: https://www.youtube.com/watch?v=-RumeaoPB2M
An example of how organizations can use Target Tracking includes managing performance for a group of EC2 Web Servers. An administrator can set a Target Tracking policy that requires CPU Utilization to remain at 70%. Should traffic increase, or decrease, CPU Utilization will remain at 70% by adding or removing EC2 instances as required.
Image Source: https://aws.amazon.com/blogs/aws/new-target-tracking-policies-for-ec2-auto-scaling/
Target Tracking can be enabled based upon multiple metrics. Barr explains:
You can track more than one target in a single Auto Scaling Group as long as each one references a distinct metric. Scaling will always choose the policy that drives the highest capacity. But, scaling will not take place if the metric has insufficient data.
Auto Scaling Target Tracking policy is currently available at no additional charge and supports Amazon EC2 and DynamoDB in all regions with plans to support all Auto Scaling resources in the coming months. Administrators can enable the policy through the Amazon EC2 and DynamoDB management consoles, AWS Command Line Interface (CLI) or API. Custom metrics can also be setup through the Auto Scaling API or CLI.
