This chapter is from the book
This chapter is from the book
This chapter is from the book
Auto Scaling
Auto Scaling enables you to quickly discover the scalable AWS resources that are part of your application and configure dynamic scaling in a matter of minutes. The Auto Scaling console provides a single user interface to use the automatic scaling features of multiple AWS services. It also offers recommendations to configure scaling for the scalable resources in your application.
Use Auto Scaling to automatically scale the following resources that support your application:
EC2 Auto Scaling groups
Aurora DB clusters
DynamoDB global secondary indexes
DynamoDB tables
ECS services
Spot Fleet requests
With Auto Scaling, you create a scaling plan with a set of instructions used to configure dynamic scaling for the scalable resources in your application. Auto Scaling creates target tracking scaling policies for the scalable resources in your scaling plan. Target tracking scaling policies adjust the capacity of your scalable resource as required to maintain resource utilization at the target value that you specified.
You can create one scaling plan per application source (an AWS CloudFormation stack or a set of tags). You can add each scalable resource to one scaling plan. If you have already configured scaling policies for a scalable resource in your application, Auto Scaling keeps the existing scaling policies instead of creating additional scaling policies for the resource.
Auto Scaling involves the creation of a Launch Configuration and an Auto Scaling group. Figure 9-2 shows the configuration of Auto Scaling in AWS.
)
Figure 9-2 Configuring Auto Scaling in AWS
Target Tracking Scaling Policies
With target tracking scaling policies, you select a predefined metric or configure a customized metric and set a target value. Application Auto Scaling creates and manages the CloudWatch alarms that trigger the scaling policy and calculates the scaling adjustment based on the metric and the target value. The scaling policy adds or removes capacity as required to keep the metric at, or close to, the specified target value. In addition to keeping the metric close to the target value, a target tracking scaling policy also adjusts to changes in the metric due to a changing load pattern and minimizes changes to the capacity of the scalable target.
When specifying a customized metric, be aware that not all metrics work for target tracking. The metric must be a valid utilization metric and describe how busy a scalable target is. The metric value must increase or decrease proportionally to the capacity of the scalable target so that the metric data can be used to proportionally scale the scalable target.
You can have multiple target tracking scaling policies for a scalable target, provided that each of them uses a different metric. Application Auto Scaling scales based on the policy that provides the largest capacity for both scale-in and scale-out. This provides greater flexibility to cover multiple scenarios and ensures that there is always enough capacity to process your application workloads.
You can also optionally disable the scale-in portion of a target tracking scaling policy. This feature provides the flexibility to use a different method for scale-in than you use for scale-out.
Keep the following in mind for Auto Scaling:
You cannot create target tracking scaling policies for Amazon EMR clusters or AppStream 2.0 fleets.
You can create 50 scaling policies per scalable target. This includes both step scaling policies and target tracking policies.
A target tracking scaling policy assumes that it should perform scale-out when the specified metric is above the target value. You cannot use a target tracking scaling policy to scale out when the specified metric is below the target value.
A target tracking scaling policy does not perform scaling when the specified metric has insufficient data. It does not perform scale-in because it does not interpret insufficient data as low utilization. To scale in when a metric has insufficient data, create a step scaling policy and have an alarm invoke the scaling policy when it changes to the INSUFFICIENT_DATA state.
You may see gaps between the target value and the actual metric data points. The reason is that Application Auto Scaling always acts conservatively by rounding up or down when it determines how much capacity to add or remove. This prevents it from adding insufficient capacity or removing too much capacity. However, for a scalable target with small capacity, the actual metric data points might seem far from the target value. For a scalable target with larger capacity, adding or removing capacity causes less of a gap between the target value and the actual metric data points.
We recommend that you scale based on metrics with a 1-minute frequency because that ensures a faster response to utilization changes. Scaling on metrics with a 5-minute frequency can result in slower response time and scaling on stale metric data.
To ensure application availability, Application Auto Scaling scales out proportionally to the metric as fast as it can but scales in more gradually.
Do not edit or delete the CloudWatch alarms that Application Auto Scaling manages for a target tracking scaling policy. Application Auto Scaling deletes the alarms automatically when you delete the Auto Scaling policy.
The Cooldown Period
The scale-out cooldown period is the amount of time, in seconds, after a scale-out activity completes before another scale-out activity can start. While this cooldown period is in effect, the capacity that has been added by the previous scale-out event that initiated the cooldown is calculated as part of the desired capacity for the next scale-out event. The intention is to continuously scale out.
The scale-in cooldown period is the amount of time, in seconds, after a scale-in activity completes before another scale-in activity can start. This cooldown period is used to block subsequent scale-in events until it has expired. The intention is to scale in conservatively to protect your application’s availability. However, if another alarm triggers a scale-out policy during the cooldown period after a scale-in event, Application Auto Scaling scales out your scalable target immediately.
