Application Auto Scaling
API Reference (API Version 2016-02-06)

PutScalingPolicy

Creates or updates a policy for an Application Auto Scaling scalable target.

Each scalable target is identified by a service namespace, resource ID, and scalable dimension. A scaling policy applies to the scalable target identified by those three attributes. You cannot create a scaling policy until you have registered the resource as a scalable target using RegisterScalableTarget.

To update a policy, specify its policy name and the parameters that you want to change. Any parameters that you don't specify are not changed by this update request.

You can view the scaling policies for a service namespace using DescribeScalingPolicies. If you are no longer using a scaling policy, you can delete it using DeleteScalingPolicy.

Multiple scaling policies can be in force at the same time for the same scalable target. You can have one or more target tracking scaling policies, one or more step scaling policies, or both. However, there is a chance that multiple policies could conflict, instructing the scalable target to scale out or in at the same time. Application Auto Scaling gives precedence to the policy that provides the largest capacity for both scale out and scale in. For example, if one policy increases capacity by 3, another policy increases capacity by 200 percent, and the current capacity is 10, Application Auto Scaling uses the policy with the highest calculated capacity (200% of 10 = 20) and scales out to 30.

Learn more about how to work with scaling policies in the Application Auto Scaling User Guide.

Request Syntax

{ "PolicyName": "string", "PolicyType": "string", "ResourceId": "string", "ScalableDimension": "string", "ServiceNamespace": "string", "StepScalingPolicyConfiguration": { "AdjustmentType": "string", "Cooldown": number, "MetricAggregationType": "string", "MinAdjustmentMagnitude": number, "StepAdjustments": [ { "MetricIntervalLowerBound": number, "MetricIntervalUpperBound": number, "ScalingAdjustment": number } ] }, "TargetTrackingScalingPolicyConfiguration": { "CustomizedMetricSpecification": { "Dimensions": [ { "Name": "string", "Value": "string" } ], "MetricName": "string", "Namespace": "string", "Statistic": "string", "Unit": "string" }, "DisableScaleIn": boolean, "PredefinedMetricSpecification": { "PredefinedMetricType": "string", "ResourceLabel": "string" }, "ScaleInCooldown": number, "ScaleOutCooldown": number, "TargetValue": number } }

Request Parameters

For information about the parameters that are common to all actions, see Common Parameters.

The request accepts the following data in JSON format.

PolicyName

The name of the scaling policy.

Type: String

Length Constraints: Minimum length of 1. Maximum length of 256.

Pattern: \p{Print}+

Required: Yes

PolicyType

The policy type. This parameter is required if you are creating a scaling policy.

The following policy types are supported:

TargetTrackingScaling—Not supported for Amazon EMR or AppStream

StepScaling—Not supported for Amazon DynamoDB

For more information, see Target Tracking Scaling Policies and Step Scaling Policies in the Application Auto Scaling User Guide.

Type: String

Valid Values: StepScaling | TargetTrackingScaling

Required: No

ResourceId

The identifier of the resource associated with the scaling policy. This string consists of the resource type and unique identifier.

  • 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.

  • EMR cluster - The resource type is instancegroup and the unique identifier is the cluster ID and instance group ID. Example: instancegroup/j-2EEZNYKUA1NTV/ig-1791Y4E1L8YI0.

  • AppStream 2.0 fleet - The resource type is fleet and the unique identifier is the fleet name. Example: fleet/sample-fleet.

  • 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.

  • Amazon SageMaker endpoint variants - The resource type is variant and the unique identifier is the resource ID. Example: endpoint/my-end-point/variant/KMeansClustering.

  • Custom resources are not supported with a resource type. This parameter must specify the OutputValue from the CloudFormation template stack used to access the resources. The unique identifier is defined by the service provider. More information is available in our GitHub repository.

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. This string consists of the service namespace, resource type, and scaling property.

  • ecs:service:DesiredCount - The desired task count of an ECS service.

  • ec2:spot-fleet-request:TargetCapacity - The target capacity of a Spot Fleet request.

  • elasticmapreduce:instancegroup:InstanceCount - The instance count of an EMR Instance Group.

  • appstream:fleet:DesiredCapacity - The desired capacity of an AppStream 2.0 fleet.

  • 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.

  • sagemaker:variant:DesiredInstanceCount - The number of EC2 instances for an Amazon SageMaker model endpoint variant.

  • custom-resource:ResourceType:Property - The scalable dimension for a custom resource provided by your own application or service.

Type: String

Valid Values: ecs:service:DesiredCount | ec2:spot-fleet-request:TargetCapacity | elasticmapreduce:instancegroup:InstanceCount | appstream:fleet:DesiredCapacity | dynamodb:table:ReadCapacityUnits | dynamodb:table:WriteCapacityUnits | dynamodb:index:ReadCapacityUnits | dynamodb:index:WriteCapacityUnits | rds:cluster:ReadReplicaCount | sagemaker:variant:DesiredInstanceCount | custom-resource:ResourceType:Property

Required: Yes

ServiceNamespace

The namespace of the AWS service that provides the resource or custom-resource for a resource provided by your own application or service. For more information, see AWS Service Namespaces in the Amazon Web Services General Reference.

Type: String

Valid Values: ecs | elasticmapreduce | ec2 | appstream | dynamodb | rds | sagemaker | custom-resource

Required: Yes

StepScalingPolicyConfiguration

A step scaling policy.

This parameter is required if you are creating a policy and the policy type is StepScaling.

Type: StepScalingPolicyConfiguration object

Required: No

TargetTrackingScalingPolicyConfiguration

A target tracking scaling policy. Includes support for predefined or customized metrics.

This parameter is required if you are creating a policy and the policy type is TargetTrackingScaling.

Type: TargetTrackingScalingPolicyConfiguration object

Required: No

Response Syntax

{ "Alarms": [ { "AlarmARN": "string", "AlarmName": "string" } ], "PolicyARN": "string" }

Response Elements

If the action is successful, the service sends back an HTTP 200 response.

The following data is returned in JSON format by the service.

Alarms

The CloudWatch alarms created for the target tracking scaling policy.

Type: Array of Alarm objects

PolicyARN

The Amazon Resource Name (ARN) of the resulting scaling policy.

Type: String

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

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

Errors

For information about the errors that are common to all actions, see Common Errors.

ConcurrentUpdateException

Concurrent updates caused an exception, for example, if you request an update to an Application Auto Scaling resource that already has a pending update.

HTTP Status Code: 400

FailedResourceAccessException

Failed access to resources caused an exception. This exception is thrown when Application Auto Scaling is unable to retrieve the alarms associated with a scaling policy due to a client error, for example, if the role ARN specified for a scalable target does not have permission to call the CloudWatch DescribeAlarms on your behalf.

HTTP Status Code: 400

InternalServiceException

The service encountered an internal error.

HTTP Status Code: 400

LimitExceededException

A per-account resource limit is exceeded. For more information, see Application Auto Scaling Limits.

HTTP Status Code: 400

ObjectNotFoundException

The specified object could not be found. For any operation that depends on the existence of a scalable target, this exception is thrown if the scalable target with the specified service namespace, resource ID, and scalable dimension does not exist. For any operation that deletes or deregisters a resource, this exception is thrown if the resource cannot be found.

HTTP Status Code: 400

ValidationException

An exception was thrown for a validation issue. Review the available parameters for the API request.

HTTP Status Code: 400

Examples

If you plan to create requests manually, you must replace the Authorization header contents in the examples (AUTHPARAMS) with a signature. For more information, see Signature Version 4 Signing Process in the Amazon Web Services General Reference. If you plan to use the AWS CLI or one of the AWS SDKs, these tools sign the requests for you.

Example of a target tracking scaling policy

The following example applies a target tracking scaling policy to an Amazon ECS service called web-app in the default cluster. The policy keeps the average CPU utilization of the service at 75 percent, with scale-out and scale-in cooldown periods of 60 seconds. The output contains the ARNs and names of the two CloudWatch alarms created on your behalf.

Sample Request

POST / HTTP/1.1 Host: autoscaling.us-west-2.amazonaws.com Accept-Encoding: identity Content-Length: 392 X-Amz-Target: AnyScaleFrontendService.PutScalingPolicy X-Amz-Date: 20160506T191044Z User-Agent: aws-cli/1.10.23 Python/2.7.11 Darwin/15.4.0 botocore/1.4.8 Content-Type: application/x-amz-json-1.1 Authorization: AUTHPARAMS { "PolicyName": "cpu75-target-tracking-scaling-policy", "PolicyType": "TargetTrackingScaling", "TargetTrackingScalingPolicyConfiguration": { "TargetValue": 75.0, "PredefinedMetricSpecification": { "PredefinedMetricType": "ECSServiceAverageCPUUtilization" }, "ScaleOutCooldown": 60, "ScaleInCooldown": 60 }, "ServiceNamespace": "ecs", "ScalableDimension": "ecs:service:DesiredCount", "ResourceId": "service/default/web-app" }

Sample Response

HTTP/1.1 200 OK x-amzn-RequestId: 4a0f8f18-cb5f-11e0-8364-37acb4b5a1b2 Content-Type: application/x-amz-json-1.1 Content-Length: 314 Date: Fri, 06 May 2016 19:10:44 GMT { "PolicyARN": "arn:aws:autoscaling:us-west-2:012345678910:scalingPolicy:6d8972f3-efc8-437c-92d1-6270f29a66e7:resource/ecs/service/default/web-app:policyName/cpu75-target-tracking-scaling-policy", "Alarms": [ { "AlarmARN": "arn:aws:cloudwatch:us-west-2:012345678910:alarm:TargetTracking-service/default/web-app-AlarmHigh-d4f0770c-b46e-434a-a60f-3b36d653feca", "AlarmName": "TargetTracking-service/default/web-app-AlarmHigh-d4f0770c-b46e-434a-a60f-3b36d653feca" }, { "AlarmARN": "arn:aws:cloudwatch:us-west-2:012345678910:alarm:TargetTracking-service/default/web-app-AlarmLow-1b437334-d19b-4a63-a812-6c67aaf2910d", "AlarmName": "TargetTracking-service/default/web-app-AlarmLow-1b437334-d19b-4a63-a812-6c67aaf2910d" } ] }

Example of a step scaling policy for scale out

The following example applies a step scaling policy to an Amazon ECS service called web-app in the default cluster. The policy increases the desired count of the service by 200%, with a cooldown period of 60 seconds. The output includes the ARN for the policy, which you need to create the CloudWatch alarm.

Sample Request

POST / HTTP/1.1 Host: autoscaling.us-west-2.amazonaws.com Accept-Encoding: identity Content-Length: 358 X-Amz-Target: AnyScaleFrontendService.PutScalingPolicy X-Amz-Date: 20160506T191138Z User-Agent: aws-cli/1.10.23 Python/2.7.11 Darwin/15.4.0 botocore/1.4.8 Content-Type: application/x-amz-json-1.1 Authorization: AUTHPARAMS { "PolicyName": "my-scaleout-policy", "PolicyType": "StepScaling", "StepScalingPolicyConfiguration": { "AdjustmentType": "PercentChangeInCapacity", "Cooldown": 60, "MetricAggregationType": "Average", "StepAdjustments": [ { "ScalingAdjustment": 200, "MetricIntervalLowerBound": 0 } ] }, "ServiceNamespace": "ecs", "ScalableDimension": "ecs:service:DesiredCount", "ResourceId": "service/default/web-app" }

Sample Response

HTTP/1.1 200 OK x-amzn-RequestId: 5ec6d08e-17ce-1e165a468-73cad4b5cel6 Content-Type: application/x-amz-json-1.1 Content-Length: 175 Date: Fri, 06 May 2016 19:11:38 GMT { "PolicyARN": "arn:aws:autoscaling:us-west-2:012345678910:scalingPolicy:ac542982-cbeb-4294-891c-a5a941dfa787:resource/ecs/service/default/web-app:policyName/my-scaleout-policy" }

Example of a step scaling policy for scale in

The following example applies a step scaling policy to the same Amazon ECS service as in the preceding example. The policy has two step adjustments that decrease the desired count of the service by 25% or 50%, depending on the size of the alarm breach, with a cooldown period of 120 seconds. The output includes the ARN for the policy, which you need to create the CloudWatch alarm.

Sample Request

POST / HTTP/1.1 Host: autoscaling.us-west-2.amazonaws.com Accept-Encoding: identity Content-Length: 431 X-Amz-Target: AnyScaleFrontendService.PutScalingPolicy X-Amz-Date: 20160506T191152Z User-Agent: aws-cli/1.10.23 Python/2.7.11 Darwin/15.4.0 botocore/1.4.8 Content-Type: application/x-amz-json-1.1 Authorization: AUTHPARAMS { "PolicyName": "my-scalein-policy", "PolicyType": "StepScaling", "StepScalingPolicyConfiguration": { "AdjustmentType": "PercentChangeInCapacity", "Cooldown": 120, "MetricAggregationType": "Average", "MinAdjustmentMagnitude": 1, "StepAdjustments": [ { "ScalingAdjustment": -25, "MetricIntervalLowerBound": -15, "MetricIntervalUpperBound": 0 }, { "ScalingAdjustment": -50, "MetricIntervalUpperBound": -15 } ] }, "ServiceNamespace": "ecs", "ScalableDimension": "ecs:service:DesiredCount", "ResourceId": "service/default/web-app" }

Sample Response

HTTP/1.1 200 OK x-amzn-RequestId: 5a64c9e1-3cfe-11e74bfad-8d1c65ec6d08 Content-Type: application/x-amz-json-1.1 Content-Length: 174 Date: Fri, 06 May 2016 19:11:52 GMT { "PolicyARN": "arn:aws:autoscaling:us-west-2:012345678910:scalingPolicy:6d8972f3-efc8-437c-92d1-6270f29a66e7:resource/ecs/service/default/web-app:policyName/my-scalein-policy" }

See Also

For more information about using this API in one of the language-specific AWS SDKs, see the following: