inference_experiments

Creates, updates, deletes or gets an inference_experiment resource or lists inference_experiments in a region

Overview

Name	inference_experiments
Type	Resource
Description	Resource Type definition for AWS::SageMaker::InferenceExperiment
Id	awscc.sagemaker.inference_experiments

Fields

get (all properties)

Name	Datatype	Description
arn	string	The Amazon Resource Name (ARN) of the inference experiment.
name	string	The name for the inference experiment.
type	string	The type of the inference experiment that you want to run.
description	string	The description of the inference experiment.
role_arn	string	The Amazon Resource Name (ARN) of an IAM role that Amazon SageMaker can assume to access model artifacts and container images, and manage Amazon SageMaker Inference endpoints for model deployment.
endpoint_name	string	The name of the endpoint used to run the monitoring job.
▶endpoint_metadata	object	The metadata of the endpoint on which the inference experiment ran.
▶schedule	object	The duration for which you want the inference experiment to run.
kms_key	string	The AWS Key Management Service (AWS KMS) key that Amazon SageMaker uses to encrypt data on the storage volume attached to the ML compute instance that hosts the endpoint.
▶data_storage_config	object	The Amazon S3 location and configuration for storing inference request and response data.
▶model_variants	array	An array of ModelVariantConfig objects. Each ModelVariantConfig object in the array describes the infrastructure configuration for the corresponding variant.
▶shadow_mode_config	object	The configuration of ShadowMode inference experiment type. Use this field to specify a production variant which takes all the inference requests, and a shadow variant to which Amazon SageMaker replicates a percentage of the inference requests. For the shadow variant also specify the percentage of requests that Amazon SageMaker replicates.
▶tags	array	An array of key-value pairs to apply to this resource.
creation_time	string	The timestamp at which you created the inference experiment.
last_modified_time	string	The timestamp at which you last modified the inference experiment.
status	string	The status of the inference experiment.
status_reason	string	The error message or client-specified reason from the StopInferenceExperiment API, that explains the status of the inference experiment.
desired_state	string	The desired state of the experiment after starting or stopping operation.
region	string	AWS region.

list (identifiers only)

Name	Datatype	Description
name	string	The name for the inference experiment.
region	string	AWS region.

For more information, see AWS::SageMaker::InferenceExperiment.

Methods

Name	Resource	Accessible by	Required Params
create_resource	inference_experiments	INSERT	Name, Type, RoleArn, EndpointName, ModelVariants, region
delete_resource	inference_experiments	DELETE	Identifier, region
update_resource	inference_experiments	UPDATE	Identifier, PatchDocument, region
list_resources	inference_experiments_list_only	SELECT	region
get_resource	inference_experiments	SELECT	Identifier, region

SELECT examples

get (all properties)

Gets all properties from an individual inference_experiment.

SELECT
  region,
  arn,
  name,
  type,
  description,
  role_arn,
  endpoint_name,
  endpoint_metadata,
  schedule,
  kms_key,
  data_storage_config,
  model_variants,
  shadow_mode_config,
  tags,
  creation_time,
  last_modified_time,
  status,
  status_reason,
  desired_state
FROM awscc.sagemaker.inference_experiments
WHERE
  region = '{{ region }}' AND
  Identifier = '{{ name }}';

list (identifiers only)

Lists all inference_experiments in a region.

SELECT
  region,
  name
FROM awscc.sagemaker.inference_experiments_list_only
WHERE
  region = '{{ region }}';

INSERT example

Use the following StackQL query and manifest file to create a new inference_experiment resource, using stack-deploy.

Required Properties

/*+ create */
INSERT INTO awscc.sagemaker.inference_experiments (
  Name,
  Type,
  RoleArn,
  EndpointName,
  ModelVariants,
  region
)
SELECT
  '{{ name }}',
  '{{ type }}',
  '{{ role_arn }}',
  '{{ endpoint_name }}',
  '{{ model_variants }}',
  '{{ region }}'
RETURNING
  ErrorCode,
  EventTime,
  Identifier,
  Operation,
  OperationStatus,
  RequestToken,
  ResourceModel,
  RetryAfter,
  StatusMessage,
  TypeName
;

All Properties

/*+ create */
INSERT INTO awscc.sagemaker.inference_experiments (
  Name,
  Type,
  Description,
  RoleArn,
  EndpointName,
  Schedule,
  KmsKey,
  DataStorageConfig,
  ModelVariants,
  ShadowModeConfig,
  Tags,
  StatusReason,
  DesiredState,
  region
)
SELECT
  '{{ name }}',
  '{{ type }}',
  '{{ description }}',
  '{{ role_arn }}',
  '{{ endpoint_name }}',
  '{{ schedule }}',
  '{{ kms_key }}',
  '{{ data_storage_config }}',
  '{{ model_variants }}',
  '{{ shadow_mode_config }}',
  '{{ tags }}',
  '{{ status_reason }}',
  '{{ desired_state }}',
  '{{ region }}'
RETURNING
  ErrorCode,
  EventTime,
  Identifier,
  Operation,
  OperationStatus,
  RequestToken,
  ResourceModel,
  RetryAfter,
  StatusMessage,
  TypeName
;

Manifest

version: 1
name: stack name
description: stack description
providers:
- aws
globals:
- name: region
  value: '{{ vars.AWS_REGION }}'
resources:
- name: inference_experiment
  props:
    - name: name
      value: '{{ name }}'
    - name: type
      value: '{{ type }}'
    - name: description
      value: '{{ description }}'
    - name: role_arn
      value: '{{ role_arn }}'
    - name: endpoint_name
      value: '{{ endpoint_name }}'
    - name: schedule
      value:
        start_time: '{{ start_time }}'
        end_time: '{{ end_time }}'
    - name: kms_key
      value: '{{ kms_key }}'
    - name: data_storage_config
      value:
        destination: '{{ destination }}'
        kms_key: '{{ kms_key }}'
        content_type:
          csv_content_types:
            - '{{ csv_content_types[0] }}'
          json_content_types:
            - '{{ json_content_types[0] }}'
    - name: model_variants
      value:
        - model_name: '{{ model_name }}'
          variant_name: '{{ variant_name }}'
          infrastructure_config:
            infrastructure_type: '{{ infrastructure_type }}'
            real_time_inference_config:
              instance_type: '{{ instance_type }}'
              instance_count: '{{ instance_count }}'
    - name: shadow_mode_config
      value:
        source_model_variant_name: '{{ source_model_variant_name }}'
        shadow_model_variants:
          - shadow_model_variant_name: '{{ shadow_model_variant_name }}'
            sampling_percentage: '{{ sampling_percentage }}'
    - name: tags
      value:
        - value: '{{ value }}'
          key: '{{ key }}'
    - name: status_reason
      value: '{{ status_reason }}'
    - name: desired_state
      value: '{{ desired_state }}'

UPDATE example

Use the following StackQL query and manifest file to update a inference_experiment resource, using stack-deploy.

/*+ update */
UPDATE awscc.sagemaker.inference_experiments
SET PatchDocument = string('{{ {
    "Description": description,
    "Schedule": schedule,
    "DataStorageConfig": data_storage_config,
    "ModelVariants": model_variants,
    "ShadowModeConfig": shadow_mode_config,
    "Tags": tags,
    "StatusReason": status_reason,
    "DesiredState": desired_state
} | generate_patch_document }}')
WHERE
  region = '{{ region }}' AND
  Identifier = '{{ name }}'
RETURNING
  ErrorCode,
  EventTime,
  Identifier,
  Operation,
  OperationStatus,
  RequestToken,
  ResourceModel,
  RetryAfter,
  StatusMessage,
  TypeName
;

DELETE example

/*+ delete */
DELETE FROM awscc.sagemaker.inference_experiments
WHERE
  Identifier = '{{ name }}' AND
  region = '{{ region }}'
RETURNING
  ErrorCode,
  EventTime,
  Identifier,
  Operation,
  OperationStatus,
  RequestToken,
  ResourceModel,
  RetryAfter,
  StatusMessage,
  TypeName
;

Additional Parameters

Mutable resources in the Cloud Control provider support additional optional parameters which can be supplied with INSERT, UPDATE, or DELETE operations. These include:

Parameter	Description
ClientToken	A unique identifier to ensure the idempotency of the resource request. This allows the provider to accurately distinguish between retries and new requests. A client token is valid for 36 hours once used. After that, a resource request with the same client token is treated as a new request. If you do not specify a client token, one is generated for inclusion in the request.
RoleArn	The ARN of the IAM role used to perform this resource operation. The role specified must have the permissions required for this operation. If you do not specify a role, a temporary session is created using your AWS user credentials.
TypeVersionId	For private resource types, the type version to use in this resource operation. If you do not specify a resource version, the default version is used.

Permissions

To operate on the inference_experiments resource, the following permissions are required:

Create

sagemaker:CreateInferenceExperiment,
sagemaker:DescribeInferenceExperiment,
sagemaker:AddTags,
sagemaker:ListTags,
iam:PassRole

Delete

sagemaker:DeleteInferenceExperiment,
sagemaker:DescribeInferenceExperiment,
sagemaker:StopInferenceExperiment,
sagemaker:ListTags

List

sagemaker:ListInferenceExperiments

Read

sagemaker:DescribeInferenceExperiment,
sagemaker:ListTags

Update

sagemaker:UpdateInferenceExperiment,
sagemaker:StartInferenceExperiment,
sagemaker:StopInferenceExperiment,
sagemaker:DescribeInferenceExperiment,
sagemaker:AddTags,
sagemaker:DeleteTags,
sagemaker:ListTags