Azure AI Search Issue - Azure AI Foundry - V1

Yogesh Dhavale (LTIMindtree Limited) 20 Reputation points Microsoft External Staff
2026-05-29T10:55:40.41+00:00

Hello ,

I am facing an issue with Azure Ai search service .Please find the attached screenshot User's image

Azure AI Search
Azure AI Search

An Azure search service with built-in artificial intelligence capabilities that enrich information to help identify and explore relevant content at scale.

0 comments

2 answers

Sort by: Most helpful
Most helpful Newest Oldest
  1. Thanmayi Godithi 11,550 Reputation points Microsoft External Staff Moderator
    2026-07-04T12:09:36.73+00:00

    Hi Yogesh Dhavale (LTIMindtree Limited),

    The AI Search connection uses key-based (API key) authentication while the search service is on a private network — that combination isn't supported for the agent tool, so the agent runtime can't resolve the private host (hence "DNS returned no data").

    Fix:

    1. In your Foundry project, open the Azure AI Search connection the agent uses and switch its auth from API key to Microsoft Entra – project managed identity (keyless).
    2. Grant the Foundry project's managed identity these roles on the search service: Search Index Data Contributor and Search Service Contributor.
    3. Because the service is private, also confirm the network path: a private endpoint on the search service, with the privatelink.search.windows.net Private DNS zone linked to the VNet your agent runtime uses (an unlinked/missing DNS zone is literally what produces "DNS returned no data"), and that your Foundry Agent Service is network-secured / VNet-injected to reach it. A public-only agent can't reach a private-only search service regardless of auth.
    4. Save the connection and re-run the agent in the playground.

    If the error is actually a 401/403 ("tool can't access the index"), that's RBAC (assign the two roles above) or a disabled API key; if it's "index not found," check the index name (case-sensitive) and that the connection points to the right search resource; if you hit "Workspace not found" while creating the connection, use az cognitiveservices account project connection create (the older az ml path doesn't work with new Foundry projects).

    Kindly let us know if the above helps or you need further assistance on this issue.

    If the answer is helpful, please click "Accept Answer" and kindly upvote it. If you have extra questions about this answer, please click "Comment".

    Was this answer helpful?

    0 comments

  2. AI answer

    2026-05-29T10:55:57.5933333+00:00

    The error in the screenshot indicates that the agent in Azure AI Foundry cannot reach the Azure AI Search resource and shows a DNS error:

    Unable to connect to Azure AI Search Resource… [DNS server returned answer with no data]

    In the documented troubleshooting for connecting Azure AI Search to Foundry agents, this specific error occurs when the Azure AI Search connection uses key-based authentication while the search service is secured with a private virtual network. Key-based authentication is not supported with private networking for agents.

    To resolve this:

    1. Open the Azure AI Foundry project and locate the Azure AI Search connection used by the agent.
    2. Change the authentication method from API key (key-based) to Microsoft Entra project managed identity (keyless authentication).
    3. Ensure the Foundry project’s managed identity has the required RBAC roles on the search service, typically:
      • Search Index Data Contributor
      • Search Service Contributor
    4. Save the updated connection and re-run the agent in the playground.

    After switching to managed-identity (keyless) auth and confirming RBAC, the agent should be able to resolve and connect to the Azure AI Search endpoint even when the service is on a private network.

    References:

    AI-generated content may be incorrect. Read our transparency notes for more information.

    Was this answer helpful?

Your answer

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