Share via

Issue with Model deployed over AML

Sudheer Madduru, Naga 0 Reputation points
2024-05-20T12:44:59.9166667+00:00

Hi,

This is regarding an issue we are facing with a deployed model over Azure ML service as described below:

  1. ML Model is built and deployed over Azure ML Real Time endpoint.
  2. we are using "HTTP" action in Power Automate to connect the ML service using REST API. We are passing the URL of the file uploaded in Azure blob storage. Once the ML service ran successfully, the tags will be generated and this output is being consumed by the Power Automate
  3. Currently, This works well for one member, but for other two it generates a fixed tag at times for all the inputs and works at times, so it's operating intermittently

Please share your ideas/thoughts to overcome the issue.  

Regards,

Sudheer

Azure Machine Learning
Azure Machine Learning
An Azure machine learning service for building and deploying models.
2,704 questions

1 answer

Sort by: Most helpful
Most helpful Newest Oldest
  1. dupammi 7,955 Reputation points Microsoft Vendor
    2024-05-21T07:56:16.7666667+00:00

    Hi @Sudheer Madduru, Naga

    Thank you for using the Microsoft Q&A forum.

    The issue with your deployed model generating intermittent fixed tags for different users could stem from several areas. First, ensure that the model is properly trained and validated to handle varied inputs. Check the REST API configuration to ensure it correctly processes the input URLs from Azure Blob Storage. Additionally, verify that Power Automate is consistently sending the correct input and properly handling the ML service's output. Finally, examine if there are any differences in the environment or permissions for the users' experiencing issues. Please share any error logs or specific examples to assist in further troubleshooting.

    0 comments