ScalingInstruction

Describes a scaling instruction for a scalable resource in a scaling plan. Each scaling instruction applies to one resource.

AWS Auto Scaling creates target tracking scaling policies based on the scaling instructions. Target tracking scaling policies adjust the capacity of your scalable resource as required to maintain resource utilization at the target value that you specified.

AWS Auto Scaling also configures predictive scaling for your Amazon EC2 Auto Scaling groups using a subset of parameters, including the load metric, the scaling metric, the target value for the scaling metric, the predictive scaling mode (forecast and scale or forecast only), and the desired behavior when the forecast capacity exceeds the maximum capacity of the resource. With predictive scaling, AWS Auto Scaling generates forecasts with traffic predictions for the two days ahead and schedules scaling actions that proactively add and remove resource capacity to match the forecast.

Important

We recommend waiting a minimum of 24 hours after creating an Auto Scaling group to configure predictive scaling. At minimum, there must be 24 hours of historical data to generate a forecast. For more information, see Best practices for scaling plans in the Scaling Plans User Guide.

MaxCapacity

The maximum capacity of the resource. The exception to this upper limit is if you specify a non-default setting for PredictiveScalingMaxCapacityBehavior.

Type: Integer

Required: Yes

MinCapacity

The minimum capacity of the resource.

Type: Integer

Required: Yes

ResourceId

The ID of the resource. This string consists of the resource type and unique identifier.

Auto Scaling group - The resource type is autoScalingGroup and the unique identifier is the name of the Auto Scaling group. Example: autoScalingGroup/my-asg.
ECS service - The resource type is service and the unique identifier is the cluster name and service name. Example: service/default/sample-webapp.
Spot Fleet request - The resource type is spot-fleet-request and the unique identifier is the Spot Fleet request ID. Example: spot-fleet-request/sfr-73fbd2ce-aa30-494c-8788-1cee4EXAMPLE.
DynamoDB table - The resource type is table and the unique identifier is the resource ID. Example: table/my-table.
DynamoDB global secondary index - The resource type is index and the unique identifier is the resource ID. Example: table/my-table/index/my-table-index.
Aurora DB cluster - The resource type is cluster and the unique identifier is the cluster name. Example: cluster:my-db-cluster.

Type: String

Length Constraints: Minimum length of 1. Maximum length of 1600.

Pattern: [\u0020-\uD7FF\uE000-\uFFFD\uD800\uDC00-\uDBFF\uDFFF\r\n\t]*

Required: Yes

ScalableDimension

The scalable dimension associated with the resource.

autoscaling:autoScalingGroup:DesiredCapacity - The desired capacity of an Auto Scaling group.
ecs:service:DesiredCount - The desired task count of an ECS service.
ec2:spot-fleet-request:TargetCapacity - The target capacity of a Spot Fleet request.
dynamodb:table:ReadCapacityUnits - The provisioned read capacity for a DynamoDB table.
dynamodb:table:WriteCapacityUnits - The provisioned write capacity for a DynamoDB table.
dynamodb:index:ReadCapacityUnits - The provisioned read capacity for a DynamoDB global secondary index.
dynamodb:index:WriteCapacityUnits - The provisioned write capacity for a DynamoDB global secondary index.
rds:cluster:ReadReplicaCount - The count of Aurora Replicas in an Aurora DB cluster. Available for Aurora MySQL-compatible edition and Aurora PostgreSQL-compatible edition.

Type: String

Required: Yes

ServiceNamespace

The namespace of the AWS service.

Type: String

Valid Values: autoscaling | ecs | ec2 | rds | dynamodb

Required: Yes

TargetTrackingConfigurations

The target tracking configurations (up to 10). Each of these structures must specify a unique scaling metric and a target value for the metric.

Type: Array of TargetTrackingConfiguration objects

Required: Yes

CustomizedLoadMetricSpecification

The customized load metric to use for predictive scaling. This parameter or a PredefinedLoadMetricSpecification is required when configuring predictive scaling, and cannot be used otherwise.

Type: CustomizedLoadMetricSpecification object

Required: No

DisableDynamicScaling

Controls whether dynamic scaling is disabled. When dynamic scaling is enabled, AWS Auto Scaling creates target tracking scaling policies based on the specified target tracking configurations.

The default is enabled (false).

Type: Boolean

Required: No

PredefinedLoadMetricSpecification

The predefined load metric to use for predictive scaling. This parameter or a CustomizedLoadMetricSpecification is required when configuring predictive scaling, and cannot be used otherwise.

Type: PredefinedLoadMetricSpecification object

Required: No

PredictiveScalingMaxCapacityBehavior

Defines the behavior that should be applied if the forecast capacity approaches or exceeds the maximum capacity specified for the resource. The default value is SetForecastCapacityToMaxCapacity.

The following are possible values:

SetForecastCapacityToMaxCapacity - AWS Auto Scaling cannot scale resource capacity higher than the maximum capacity. The maximum capacity is enforced as a hard limit.
SetMaxCapacityToForecastCapacity - AWS Auto Scaling may scale resource capacity higher than the maximum capacity to equal but not exceed forecast capacity.
SetMaxCapacityAboveForecastCapacity - AWS Auto Scaling may scale resource capacity higher than the maximum capacity by a specified buffer value. The intention is to give the target tracking scaling policy extra capacity if unexpected traffic occurs.

Only valid when configuring predictive scaling.

Type: String

Valid Values: SetForecastCapacityToMaxCapacity | SetMaxCapacityToForecastCapacity | SetMaxCapacityAboveForecastCapacity

Required: No

PredictiveScalingMaxCapacityBuffer

The size of the capacity buffer to use when the forecast capacity is close to or exceeds the maximum capacity. The value is specified as a percentage relative to the forecast capacity. For example, if the buffer is 10, this means a 10 percent buffer, such that if the forecast capacity is 50, and the maximum capacity is 40, then the effective maximum capacity is 55.

Only valid when configuring predictive scaling. Required if the PredictiveScalingMaxCapacityBehavior is set to SetMaxCapacityAboveForecastCapacity, and cannot be used otherwise.

The range is 1-100.

Type: Integer

Required: No

PredictiveScalingMode

The predictive scaling mode. The default value is ForecastAndScale. Otherwise, AWS Auto Scaling forecasts capacity but does not create any scheduled scaling actions based on the capacity forecast.

Type: String

Valid Values: ForecastAndScale | ForecastOnly

Required: No

ScalingPolicyUpdateBehavior

Controls whether a resource's externally created scaling policies are kept or replaced.

The default value is KeepExternalPolicies. If the parameter is set to ReplaceExternalPolicies, any scaling policies that are external to the scaling plan are deleted and new target tracking scaling policies created.

Only valid when configuring dynamic scaling.

Condition: The number of existing policies to be replaced must be less than or equal to 50. If there are more than 50 policies to be replaced, AWS Auto Scaling keeps all existing policies and does not create new ones.

Type: String

Valid Values: KeepExternalPolicies | ReplaceExternalPolicies

Required: No

ScheduledActionBufferTime

The amount of time, in seconds, to buffer the run time of scheduled scaling actions when scaling out. For example, if the forecast says to add capacity at 10:00 AM, and the buffer time is 5 minutes, then the run time of the corresponding scheduled scaling action will be 9:55 AM. The intention is to give resources time to be provisioned. For example, it can take a few minutes to launch an EC2 instance. The actual amount of time required depends on several factors, such as the size of the instance and whether there are startup scripts to complete.

The value must be less than the forecast interval duration of 3600 seconds (60 minutes). The default is 300 seconds.

Only valid when configuring predictive scaling.

Type: Integer

Valid Range: Minimum value of 0.

Required: No

ScalingInstruction

Important

Contents

See Also