Amazon Web Service introduced new Auto Scaling model that consist with a very vital feature which is called Target Tracking Policy.
Auto Scaling Policy is created to make use of the target tracking policy where you will need to select a target value for a particular CloudWatch metric. Auto Scaling then processes the metric toward the target at the same time adjusting the applicable CloudWatch Alarms. Choose the desired target according to your preferred metrics-driven units for your application which will be convenient and forthright instead of setting up thresholds and ranges manually by utilizing the original step scaling policy type.
Target Tracking can use used in conjunction with step scaling so that to apply an advanced scaling strategy. You can use the Target Tracking for scale-out operation and step scaling for scale-in.
AWS has added target tracking support to Elastic Compute Cloud Auto Scaling. Scaling Policies can now be created which are driven by CPU load, application load balancer request counts, custom metric or network traffic. The Request Count per Target metric is newly launched and is also part of the Target Tracking Policy.
Application load balancer Request Count Per target, Average CPU Utilization, Average Network in (bytes) and Average Network out (bytes) are the metric type which shares an important property thus adding additional Elastic Compute Cloud instance will drive the metric down and vice versa with no changes in the overall load.
Auto Scaling group can be created that will make use of the target tracking you have to simply enter the name for the policy, select a metric and choose the desired target value. There is also an option to disable the scale-in side of the policy. By doing this you can scale in using a separate policy or you can do this manually. The target tracking policies can be created by using the AWS Command Line Interface (CLI), AWS SDKs or the AWS Management Console.
The following are point that should be kept in mind while using the target tracking:-
1. In Auto Scaling Group you can track more than one target as long as each reference is a distinct metric. Scaling will always select the policy that delivers the highest capacity.
2. If the metric has insufficient data then the Scaling will not take place.
3. Auto Scaling adjusts for transient rapid fluctuations in the metrics and striving to minimize the corresponding fluctuations in capacity.
4. Target Tracking can be set up for a custom metric via the AWS Command Line Interface (CLI) or Auto Scaling API.
5. In the most cases, it is suggested to choose to scale on metrics that is launched with a 1-minute frequency which is also known as the detailed monitoring. There will be slower response time by using the 5-minute metrics as the basis for scaling.
No comments:
Post a Comment