Unable to Use Grounding with Bing Search in Private Endpoint–Only Azure AI Foundry Projects

木本 瑞希 80 Reputation points
2025-06-12T09:07:41.55+00:00

We are trying to create an AI agent with Grounding with Bing Search in an Azure AI Foundry project. However, when the Foundry project is configured to be accessible only via private endpoint, we encounter the following error when attempting to connect the knowledge source (Grounding with Bing Search):

Failed to create connection   NotFoundError: Unable to find workspace: /subscriptions/<subscription ID>/resourcegroups/<resource group name>/providers/Microsoft.MachineLearningServices/workspaces/<project name>

It seems the project cannot access the required workspace due to network restrictions.

Is it currently not possible to use knowledge-grounded agents (Grounding with Bing Search) within a Foundry project that is restricted to private endpoints only?

Any clarification or recommended workaround would be appreciated.

Azure AI services
Azure AI services
A group of Azure services, SDKs, and APIs designed to make apps more intelligent, engaging, and discoverable.
3,632 questions
{count} votes

Accepted answer
  1. Saideep Anchuri 9,500 Reputation points Moderator
    2025-06-13T01:06:40.0733333+00:00

    Hi 木本 瑞希

    Currently, using knowledge-grounded agents like Grounding with Bing Search within an Azure AI Foundry project that is restricted to private endpoints may present challenges. The error you are encountering suggests that the project cannot access the required workspace due to network restrictions imposed by the private endpoint configuration and most probably DNS resolution is not happening for foundry hub project workspace.

    As per the context, it is important to ensure that the necessary roles and permissions are correctly assigned and that the networking settings allow for the required connections. Specifically, if the Azure AI Foundry project is set to be accessible only via private endpoints, you need to ensure that the private endpoint connections are properly established and that the Azure resources involved are configured to allow access from the private network.

    Kindly refer below link: setup

    end-to-end-secured-networking-for-agent-service

    https://learn.microsoft.com/en-us/azure/ai-foundry/how-to/troubleshoot-secure-connection-project

    Thank You.

    0 comments No comments

1 additional answer

Sort by: Most helpful
  1. Danny Dang 90 Reputation points Independent Advisor
    2025-06-17T09:45:04.9333333+00:00

    Hi Mizuki,

    Thank you for contacting Q&A Forum.

    It looks like the issue you're encountering with using Grounding with Bing Search in a private endpoint-only Azure AI Foundry project is related to network configuration. According to the Azure AI Services documentation, you need to ensure that your virtual network (VNet) and private endpoint (PE) settings are correctly configured to allow the necessary connections.

    Here are some steps to help you troubleshoot and resolve the issue:

    1. Configure Virtual Network (VNet): Ensure that your VNet is properly set up to allow communication between your Azure AI Foundry project and the required Azure resources.
    2. Set Up Private Endpoints (PE): Make sure that the private endpoints are correctly configured to enable access to the necessary services, such as the workspace for Grounding with Bing Search.
    3. Network Security Groups (NSGs): Check your NSG rules to ensure that they allow the required traffic between your VNet and the Azure services.
    4. DNS Configuration: Verify that your DNS settings are correctly configured to resolve the necessary service endpoints.

    For detailed instructions, you can refer to the VNet and PE configuration guide. https://learn.microsoft.com/en-us/azure/ai-services/agents/how-to/virtual-networks

    If I have answered your question, please accept this answer as a token of appreciation and don't forget to give a thumbs up for "Was it helpful"!

    Best regards,

    0 comments No comments

Your answer

Answers can be marked as Accepted Answers by the question author, which helps users to know the answer solved the author's problem.