clusters

Creates, updates, deletes, gets or lists a clusters resource.

Overview

Name	clusters
Type	Resource
Id	aws.sagemaker.clusters

Fields

The following fields are returned by SELECT queries:

describe_cluster

Name	Datatype	Description
auto_scaling	object	The current autoscaling configuration and status for the autoscaler.
cluster_arn	string	The Amazon Resource Name (ARN) of the SageMaker HyperPod cluster. (pattern: <code>arn:aws[a-z-]:sagemaker:[a-z0-9-]:[0-9]{12}:cluster/[a-z0-9]{12}</code>)
cluster_name	string	The name of the SageMaker HyperPod cluster. (pattern: <code>[a-zA-Z0-9](-[a-zA-Z0-9])</code>)
cluster_role	string	The Amazon Resource Name (ARN) of the IAM role that HyperPod uses for cluster autoscaling operations. (pattern: <code>arn:aws[a-z-]*:iam::\d{12}:role/?[a-zA-Z_0-9+=,.@-_/]+</code>)
cluster_status	string	The status of the SageMaker HyperPod cluster. (Creating, Deleting, Failed, InService, RollingBack, SystemUpdating, Updating)
creation_time	string (date-time)	The time when the SageMaker Cluster is created.
failure_message	string	The failure message of the SageMaker HyperPod cluster.
instance_groups	array	The instance groups of the SageMaker HyperPod cluster.
node_provisioning_mode	string	The mode used for provisioning nodes in the cluster. (Continuous)
node_recovery	string	The node recovery mode configured for the SageMaker HyperPod cluster. (Automatic, None)
orchestrator	object	The type of orchestrator used for the SageMaker HyperPod cluster.
restricted_instance_groups	array	The specialized instance groups for training models like Amazon Nova to be created in the SageMaker HyperPod cluster.
restricted_instance_groups_config	object	The configuration for the restricted instance groups (RIG) in the SageMaker HyperPod cluster.
tiered_storage_config	object	The current configuration for managed tier checkpointing on the HyperPod cluster. For example, this shows whether the feature is enabled and the percentage of cluster memory allocated for checkpoint storage.
vpc_config	object	Specifies an Amazon Virtual Private Cloud (VPC) that your SageMaker jobs, hosted models, and compute resources have access to. You can control access to and from your resources by configuring a VPC. For more information, see Give SageMaker Access to Resources in your Amazon VPC.

list_clusters

Name	Datatype	Description
cluster_arn	string	The Amazon Resource Name (ARN) of the SageMaker HyperPod cluster. (pattern: <code>arn:aws[a-z-]:sagemaker:[a-z0-9-]:[0-9]{12}:cluster/[a-z0-9]{12}</code>)
cluster_name	string	The name of the SageMaker HyperPod cluster. (pattern: <code>[a-zA-Z0-9](-[a-zA-Z0-9])</code>)
cluster_status	string	The status of the SageMaker HyperPod cluster. (Creating, Deleting, Failed, InService, RollingBack, SystemUpdating, Updating)
creation_time	string (date-time)	The time when the SageMaker HyperPod cluster is created.
image_version_status	string	The aggregate status of the image version across the cluster's instance groups. (UpToDate, UpdateAvailable, SecurityUpdateRequired, EndOfLife)
training_plan_arns	array	A list of Amazon Resource Names (ARNs) of the training plans associated with this cluster. For more information about how to reserve GPU capacity for your SageMaker HyperPod clusters using Amazon SageMaker Training Plan, see CreateTrainingPlan .

Methods

The following methods are available for this resource:

Name	Accessible by	Required Params	Optional Params	Description
describe_cluster	select	region		Retrieves information of a SageMaker HyperPod cluster.
list_clusters	select	region		Retrieves the list of SageMaker HyperPod clusters.
create_cluster	insert	region, ClusterName		Creates an Amazon SageMaker HyperPod cluster. SageMaker HyperPod is a capability of SageMaker for creating and managing persistent clusters for developing large machine learning models, such as large language models (LLMs) and diffusion models. To learn more, see Amazon SageMaker HyperPod in the Amazon SageMaker Developer Guide.
update_cluster	update	region, ClusterName		Updates a SageMaker HyperPod cluster.
delete_cluster	delete	region		Delete a SageMaker HyperPod cluster.
start_cluster_health_check	exec	region, ClusterName, DeepHealthCheckConfigurations		Start deep health checks for a SageMaker HyperPod cluster. You can use DescribeClusterNode API to track progress of the deep health checks. The unhealthy nodes will be automatically rebooted or replaced. Please see Resilience-related Kubernetes labels by SageMaker HyperPod for details.
update_cluster_software	exec	region, ClusterName		Updates the platform software of a SageMaker HyperPod cluster for security patching. To learn how to use this API, see Update the SageMaker HyperPod platform software of a cluster. The UpgradeClusterSoftware API call may impact your SageMaker HyperPod cluster uptime and availability. Plan accordingly to mitigate potential disruptions to your workloads.

Parameters

Parameters can be passed in the WHERE clause of a query. Check the Methods section to see which parameters are required or optional for each operation.

Name	Datatype	Description
region	string	AWS region (default: us-east-1)

SELECT examples

describe_cluster

Retrieves information of a SageMaker HyperPod cluster.

SELECT
auto_scaling,
cluster_arn,
cluster_name,
cluster_role,
cluster_status,
creation_time,
failure_message,
instance_groups,
node_provisioning_mode,
node_recovery,
orchestrator,
restricted_instance_groups,
restricted_instance_groups_config,
tiered_storage_config,
vpc_config
FROM aws.sagemaker.clusters
WHERE region = '{{ region }}' -- required
;

list_clusters

Retrieves the list of SageMaker HyperPod clusters.

SELECT
cluster_arn,
cluster_name,
cluster_status,
creation_time,
image_version_status,
training_plan_arns
FROM aws.sagemaker.clusters
WHERE region = '{{ region }}' -- required
;

INSERT examples

create_cluster

Creates an Amazon SageMaker HyperPod cluster. SageMaker HyperPod is a capability of SageMaker for creating and managing persistent clusters for developing large machine learning models, such as large language models (LLMs) and diffusion models. To learn more, see Amazon SageMaker HyperPod in the Amazon SageMaker Developer Guide.

INSERT INTO aws.sagemaker.clusters (
ClusterName,
InstanceGroups,
RestrictedInstanceGroups,
RestrictedInstanceGroupsConfig,
VpcConfig,
Tags,
Orchestrator,
NodeRecovery,
TieredStorageConfig,
NodeProvisioningMode,
ClusterRole,
AutoScaling,
region
)
SELECT 
'{{ ClusterName }}' /* required */,
'{{ InstanceGroups }}',
'{{ RestrictedInstanceGroups }}',
'{{ RestrictedInstanceGroupsConfig }}',
'{{ VpcConfig }}',
'{{ Tags }}',
'{{ Orchestrator }}',
'{{ NodeRecovery }}',
'{{ TieredStorageConfig }}',
'{{ NodeProvisioningMode }}',
'{{ ClusterRole }}',
'{{ AutoScaling }}',
'{{ region }}'
RETURNING
cluster_arn
;

Manifest

# Description fields are for documentation purposes
- name: clusters
props:
  - name: region
    value: "{{ region }}"
    description: Required parameter for the clusters resource.
  - name: ClusterName
    value: "{{ ClusterName }}"
    description: |
      The name for the new SageMaker HyperPod cluster.
  - name: InstanceGroups
    description: |
      The instance groups to be created in the SageMaker HyperPod cluster.
    value:
      - InstanceCount: {{ InstanceCount }}
        MinInstanceCount: {{ MinInstanceCount }}
        InstanceGroupName: "{{ InstanceGroupName }}"
        InstanceType: "{{ InstanceType }}"
        InstanceRequirements:
          InstanceTypes:
            - "{{ InstanceTypes }}"
        LifeCycleConfig:
          SourceS3Uri: "{{ SourceS3Uri }}"
          OnCreate: "{{ OnCreate }}"
          OnInitComplete: "{{ OnInitComplete }}"
        ExecutionRole: "{{ ExecutionRole }}"
        ThreadsPerCore: {{ ThreadsPerCore }}
        InstanceStorageConfigs: "{{ InstanceStorageConfigs }}"
        OnStartDeepHealthChecks: "{{ OnStartDeepHealthChecks }}"
        TrainingPlanArn: "{{ TrainingPlanArn }}"
        OverrideVpcConfig:
          SecurityGroupIds:
            - "{{ SecurityGroupIds }}"
          Subnets:
            - "{{ Subnets }}"
        ScheduledUpdateConfig:
          ScheduleExpression: "{{ ScheduleExpression }}"
          DeploymentConfig:
            RollingUpdatePolicy:
              MaximumBatchSize:
                Type: "{{ Type }}"
                Value: {{ Value }}
              RollbackMaximumBatchSize:
                Type: "{{ Type }}"
                Value: {{ Value }}
            WaitIntervalInSeconds: {{ WaitIntervalInSeconds }}
            AutoRollbackConfiguration:
              - AlarmName: "{{ AlarmName }}"
        ImageId: "{{ ImageId }}"
        AutoPatchConfig:
          PatchingStrategy: "{{ PatchingStrategy }}"
          PatchSchedule:
            NextPatchDate: "{{ NextPatchDate }}"
          DeploymentConfig:
            RollingUpdatePolicy:
              MaximumBatchSize:
                Type: "{{ Type }}"
                Value: {{ Value }}
              RollbackMaximumBatchSize:
                Type: "{{ Type }}"
                Value: {{ Value }}
            WaitIntervalInSeconds: {{ WaitIntervalInSeconds }}
            AutoRollbackConfiguration:
              - AlarmName: "{{ AlarmName }}"
        ImageReleaseVersion: "{{ ImageReleaseVersion }}"
        KubernetesConfig:
          Labels: "{{ Labels }}"
          Taints:
            - Key: "{{ Key }}"
              Value: "{{ Value }}"
              Effect: "{{ Effect }}"
        SlurmConfig:
          NodeType: "{{ NodeType }}"
          PartitionNames:
            - "{{ PartitionNames }}"
        CapacityRequirements:
          Spot: "{{ Spot }}"
          OnDemand: "{{ OnDemand }}"
        NetworkInterface:
          InterfaceType: "{{ InterfaceType }}"
  - name: RestrictedInstanceGroups
    description: |
      The specialized instance groups for training models like Amazon Nova to be created in the SageMaker HyperPod cluster.
    value:
      - InstanceCount: {{ InstanceCount }}
        InstanceGroupName: "{{ InstanceGroupName }}"
        InstanceType: "{{ InstanceType }}"
        ExecutionRole: "{{ ExecutionRole }}"
        ThreadsPerCore: {{ ThreadsPerCore }}
        InstanceStorageConfigs: "{{ InstanceStorageConfigs }}"
        OnStartDeepHealthChecks: "{{ OnStartDeepHealthChecks }}"
        TrainingPlanArn: "{{ TrainingPlanArn }}"
        OverrideVpcConfig:
          SecurityGroupIds:
            - "{{ SecurityGroupIds }}"
          Subnets:
            - "{{ Subnets }}"
        ScheduledUpdateConfig:
          ScheduleExpression: "{{ ScheduleExpression }}"
          DeploymentConfig:
            RollingUpdatePolicy:
              MaximumBatchSize:
                Type: "{{ Type }}"
                Value: {{ Value }}
              RollbackMaximumBatchSize:
                Type: "{{ Type }}"
                Value: {{ Value }}
            WaitIntervalInSeconds: {{ WaitIntervalInSeconds }}
            AutoRollbackConfiguration:
              - AlarmName: "{{ AlarmName }}"
        EnvironmentConfig:
          FSxLustreConfig:
            SizeInGiB: {{ SizeInGiB }}
            PerUnitStorageThroughput: {{ PerUnitStorageThroughput }}
  - name: RestrictedInstanceGroupsConfig
    description: |
      The configuration for the restricted instance groups (RIG) in the SageMaker HyperPod cluster.
    value:
      SharedEnvironmentConfig:
        FSxLustreDeletionPolicy: "{{ FSxLustreDeletionPolicy }}"
        FSxLustreConfig:
          SizeInGiB: {{ SizeInGiB }}
          PerUnitStorageThroughput: {{ PerUnitStorageThroughput }}
  - name: VpcConfig
    description: |
      Specifies an Amazon Virtual Private Cloud (VPC) that your SageMaker jobs, hosted models, and compute resources have access to. You can control access to and from your resources by configuring a VPC. For more information, see Give SageMaker Access to Resources in your Amazon VPC.
    value:
      SecurityGroupIds:
        - "{{ SecurityGroupIds }}"
      Subnets:
        - "{{ Subnets }}"
  - name: Tags
    description: |
      Custom tags for managing the SageMaker HyperPod cluster as an Amazon Web Services resource. You can add tags to your cluster in the same way you add them in other Amazon Web Services services that support tagging. To learn more about tagging Amazon Web Services resources in general, see Tagging Amazon Web Services Resources User Guide.
    value:
      - Key: "{{ Key }}"
        Value: "{{ Value }}"
  - name: Orchestrator
    description: |
      The type of orchestrator to use for the SageMaker HyperPod cluster. Currently, supported values are "Eks" and "Slurm", which is to use an Amazon Elastic Kubernetes Service or Slurm cluster as the orchestrator. If you specify the Orchestrator field, you must provide exactly one orchestrator configuration: either Eks or Slurm. Specifying both or providing an empty configuration returns a validation error.
    value:
      Eks:
        ClusterArn: "{{ ClusterArn }}"
      Slurm:
        SlurmConfigStrategy: "{{ SlurmConfigStrategy }}"
  - name: NodeRecovery
    value: "{{ NodeRecovery }}"
    description: |
      The node recovery mode for the SageMaker HyperPod cluster. When set to Automatic, SageMaker HyperPod will automatically reboot or replace faulty nodes when issues are detected. When set to None, cluster administrators will need to manually manage any faulty cluster instances.
    valid_values: ['Automatic', 'None']
  - name: TieredStorageConfig
    description: |
      The configuration for managed tier checkpointing on the HyperPod cluster. When enabled, this feature uses a multi-tier storage approach for storing model checkpoints, providing faster checkpoint operations and improved fault tolerance across cluster nodes.
    value:
      Mode: "{{ Mode }}"
      InstanceMemoryAllocationPercentage: {{ InstanceMemoryAllocationPercentage }}
  - name: NodeProvisioningMode
    value: "{{ NodeProvisioningMode }}"
    description: |
      The mode for provisioning nodes in the cluster. You can specify the following modes: Continuous: Scaling behavior that enables 1) concurrent operation execution within instance groups, 2) continuous retry mechanisms for failed operations, 3) enhanced customer visibility into cluster events through detailed event streams, 4) partial provisioning capabilities. Your clusters and instance groups remain InService while scaling. This mode is only supported for EKS orchestrated clusters.
    valid_values: ['Continuous']
  - name: ClusterRole
    value: "{{ ClusterRole }}"
    description: |
      The Amazon Resource Name (ARN) of the IAM role that HyperPod assumes to perform cluster autoscaling operations. This role must have permissions for sagemaker:BatchAddClusterNodes and sagemaker:BatchDeleteClusterNodes. This is only required when autoscaling is enabled and when HyperPod is performing autoscaling operations.
  - name: AutoScaling
    description: |
      The autoscaling configuration for the cluster. Enables automatic scaling of cluster nodes based on workload demand using a Karpenter-based system.
    value:
      Mode: "{{ Mode }}"
      AutoScalerType: "{{ AutoScalerType }}"

UPDATE examples

update_cluster

Updates a SageMaker HyperPod cluster.

UPDATE aws.sagemaker.clusters
SET 
ClusterName = '{{ ClusterName }}',
InstanceGroups = '{{ InstanceGroups }}',
RestrictedInstanceGroups = '{{ RestrictedInstanceGroups }}',
RestrictedInstanceGroupsConfig = '{{ RestrictedInstanceGroupsConfig }}',
TieredStorageConfig = '{{ TieredStorageConfig }}',
NodeRecovery = '{{ NodeRecovery }}',
InstanceGroupsToDelete = '{{ InstanceGroupsToDelete }}',
NodeProvisioningMode = '{{ NodeProvisioningMode }}',
ClusterRole = '{{ ClusterRole }}',
AutoScaling = '{{ AutoScaling }}',
Orchestrator = '{{ Orchestrator }}'
WHERE 
region = '{{ region }}' --required
AND ClusterName = '{{ ClusterName }}' --required
RETURNING
cluster_arn;

DELETE examples

delete_cluster

Delete a SageMaker HyperPod cluster.

DELETE FROM aws.sagemaker.clusters
WHERE region = '{{ region }}' --required
;

Lifecycle Methods

start_cluster_health_check

Start deep health checks for a SageMaker HyperPod cluster. You can use DescribeClusterNode API to track progress of the deep health checks. The unhealthy nodes will be automatically rebooted or replaced. Please see Resilience-related Kubernetes labels by SageMaker HyperPod for details.

EXEC aws.sagemaker.clusters.start_cluster_health_check 
@region='{{ region }}' --required 
@@json=
'{
"ClusterName": "{{ ClusterName }}", 
"DeepHealthCheckConfigurations": "{{ DeepHealthCheckConfigurations }}"
}'
;

update_cluster_software

Updates the platform software of a SageMaker HyperPod cluster for security patching. To learn how to use this API, see Update the SageMaker HyperPod platform software of a cluster. The UpgradeClusterSoftware API call may impact your SageMaker HyperPod cluster uptime and availability. Plan accordingly to mitigate potential disruptions to your workloads.

EXEC aws.sagemaker.clusters.update_cluster_software 
@region='{{ region }}' --required 
@@json=
'{
"ClusterName": "{{ ClusterName }}", 
"InstanceGroups": "{{ InstanceGroups }}", 
"DeploymentConfig": "{{ DeploymentConfig }}", 
"ImageId": "{{ ImageId }}"
}'
;