Hi @Daryl M, I'm not very familiar with AWS' Sagemaker, but from what you're describing, a Logic App or Power Automate Flow sounds like an appropriate solution. I would check Sagemaker REST API for an endpoint that represents the completed job that you currently have. You can leverage the HTTP action in Power Automate or Logic Apps to call your Sagemaker endpoint, providing the necessary input parameters, and process the output as necessary. You can utilize the Power BI connector to process the output of the API call. If Sagemaker supports webhooks, you can leverage HTTP trigger action in a Logic App where Sagemaker calls your Logic app indicating there is data to be processed or downloaded. Here's an example how you can use either Logic Apps or Power Automate with SageMaker:
- Create a new Logic App or Power Automate flow.
- Add a Recurrence trigger to the flow to run the flow at a specific interval.
- Add an HTTP action to the flow. In the HTTP action, specify the SageMaker CreateTrainingJob API endpoint and provide the necessary input parameters.
- Add an Azure AD action to the flow to authenticate the request.
- Add a Parse JSON action to the flow to parse the output from the SageMaker API.
- Add a Condition action to the flow to check if the training job was successfully created.
- If the training job was successfully created, add any additional actions to the flow that you need to process the output from the SageMaker API.
- If the training job was not successfully created, add an action to the flow to send an email notification or create a ticket in your ITSM system.
- Save and test the flow.