Delete a Human Review Workflow - Amazon SageMaker

Delete a Human Review Workflow

When you delete a human review workflow or you delete your AWS account while a human loop is in process, your human review workflow status changes to Deleting. Amazon A2I automatically stops and deletes all associated human loops if workers have not started tasks created by those human loops. If human workers are already working on a task, that task continues to be available until it is completed or expires. As long as workers are still working on a task, your human review workflow's status is Deleting. If these tasks are completed, the results are stored in the Amazon S3 bucket specified in your flow definition.

Deleting a flow definition does not remove any worker answers from your S3 bucket. If the tasks are completed, but you deleted your AWS account, the results are stored in the Augmented AI service bucket for 30 days and then permanently deleted.

After all human loops have been deleted, the human review workflow is permanently deleted. When a human review workflow has been deleted, you can reuse its name to create a new human review workflow.

You might want to delete a human review workflow for any of the following reasons:

  • You have sent data to a set of human reviewers and you want to delete all non-started human loops because you do not want those workers to work on those tasks any longer.

  • The worker task template used to generate your worker UI does not render correctly or is not functioning as expected.

After you delete a human review workflow, the following changes occur:

  • The human review workflow no longer appears on the Human review workflows page in the Augmented AI area of the Amazon SageMaker console.

  • When you use the human review workflow name as input to the API operations DescribeFlowDefinition or DeleteFlowDefinition, Augmented AI returns a ResourceNotFound error.

  • When you use ListFlowDefinitions, deleted human review workflows aren't included in the results.

  • When you use the human review workflow ARN as input to the Augmented AI Runtime API operation ListHumanLoops, Augmented AI returns a ResourceNotFoundException.

Delete a Flow Definition Using the Console or the SageMaker API

You can delete a human review workflow on the Human review workflows page in the Augmented AI area of the SageMaker console or by using the SageMaker API.

Flow definitions can only be deleted if their status is Active.

Delete a human review workflow (console)

  1. Navigate to the Augmented AI console at https://console.aws.amazon.com/a2i/.

  2. In the navigation pane, under the Augmented AI section, choose Human review workflows.

  3. Select the hyperlinked name of the human review workflow that you want to delete.

  4. On the Summary page of your human review workflow, choose Delete.

  5. In the dialog box asking you to confirm that you want to delete your human review workflow, choose Delete.

You're automatically redirected to the Human review workflows page. While your human review workflow is being deleted, the status Deleting appears in the status column for that workflow. After it's deleted, it doesn't appear in the list of workflows on this page.

Delete a human review workflow (API)

You can delete a human review workflow (flow definition) using the SageMaker DeleteFlowDefinition API operation. This API operation is supported through the AWS CLI and a variety of language specific SDKs. The following table shows example requests using SDK for Python (Boto3) and the AWS CLI to delete the human review workflow, example-flow-definition.

AWS SDK for Python (Boto3)

The following request example uses the SDK for Python (Boto3) to delete the human review workflow. For more information, see delete_flow_definition in the AWS SDK for Python (Boto) API Reference.

import boto3 sagemaker_client = boto3.client('sagemaker') response = sagemaker_client.delete_flow_definition(FlowDefinitionName='example-flow-definition')
AWS CLI

The following request example uses the AWS CLI to delete the human review workflow. For more information, see delete-flow-definition in the AWS CLI Command Reference.

$ aws sagemaker delete-flow-definition --flow-definition-name 'example-flow-definition'

If the action is successful, Augmented AI sends back an HTTP 200 response with an empty HTTP body.