Fabric Data Agent Tool not working

[AyusmanBasu0604]

Issue with Microsoft Foundry Agent tool: Fabric Data Agent.
Getting error:

tool_user_error: Fabric run failed during execution. Find out more troubleshooting details here - https://aka.ms/foundryfabrictroubleshooting
I have checked the troubleshooting page but nothing useful as the error is "too generic"

Foundry Agent is using Model: Gpt-4.1
Tool connected: Fabric Data Agent

My Agent was working fine till few hours back and it's been working since last few days ever since it's been created.

I have admin access for my Entra ID in Fabric workspace and also Data Agent.
I have both Azure AI User and Azure AI Admin role at foundry service and project level.
Even provided the Foundry service project's Managed Identity Admin rights on the Fabric Workspace.
I have separately tested the Fabric Data Agent from Fabric workspace and it's working absolutely fine.
I have even created a new connection for the Fabric Data Agent and tried with a new agent, still it's same.
I have also tried by changing the LLM Model.

Not sure what happened suddenly! Can anyone help or suggest?
Also, is there any Quota limit?

[v-pnaroju-msft]

Hi AyusmanBasu0604,

Thank you for the update and for sharing your insights and working approach, which will be beneficial to other members of the community. Should you have any further queries, please feel free to contact the Microsoft Fabric community.

Thank you.

[v-pnaroju-msft]

Hi AyusmanBasu0604,

We are following up to see whether what we shared resolved your issue. If the issue still persists, please raise a Microsoft Support ticket using the provided link: Microsoft Fabric Support and Status | Microsoft Fabric, so that engineering can review the backend logs.

If you need more support, please reach out to the Microsoft Fabric community.

Thank you.

[AyusmanBasu0604 · Accepted solution]

well the issue didn't get resolved and I ended up creating agents through a different Foundry Project (New Foundry Experience) and Region, it's working in that for now.  @v-pnaroju-msft 

[v-pnaroju-msft]

Thank you, @bariscihan for your response.

Hi AyusmanBasu0604,

We appreciate your inquiry on the Microsoft Fabric Community Forum.

Based on my understanding, the issue is unlikely to be related to permissions, configuration, model choice, or the Fabric Data Agent itself, since the same agent executes successfully when run directly in Microsoft Fabric and also from Classic Foundry (East US). The failure occurs only in the new Azure AI Foundry (East US2).

A temporary workaround is to continue using Classic Foundry Agents for Fabric Data Agent execution. Since this behavior started abruptly, it may be region or runtime specific. We kindly request you to raise a Microsoft Support ticket using the provided link: Microsoft Fabric Support and Status | Microsoft Fabric so that engineering can review the backend logs.

We hope the information provided will assist in resolving the issue. Should you have any further queries, please feel free to contact the Microsoft Fabric community.

Thank you.

[bariscihan]

Hi,
From the behavior described, this looks less like a Fabric Data Agent configuration issue and more like an execution or service-layer issue between Foundry Agent and Fabric.

If the Fabric Data Agent works correctly when triggered directly from the Fabric workspace, but fails only when invoked via Foundry Agent, the failure is likely happening during the tool execution handshake (Foundry → Fabric job execution). The tool_user_error: Fabric run failed during execution message is typically returned when Fabric fails before returning a detailed internal error.

Since you already validated:

• Entra ID permissions (Fabric workspace + Data Agent access)

• Azure AI User/Admin roles at Foundry level

• Managed Identity access to Fabric workspace

• New connection + new agent test

• Model switch test

• Fabric Data Agent standalone execution success

I would mainly suspect one of the following:

• Temporary backend/service-side issue (very common when errors are generic)
• Token delegation or identity refresh issue between Foundry and Fabric
• Capacity / quota / throttling scenario (especially if this started suddenly)
• Recent backend runtime update or regression on Foundry tool connector

---

Suggested Quick Checks

• Check Fabric Capacity metrics and CU utilization at failure time
• Check Fabric Activity Logs for failed runs triggered by service principal / MI
• Re-create Foundry → Fabric connection (even if already done, forces token refresh)
• Test with a minimal Data Agent query (to rule out timeout / payload issues)
• Verify Foundry project region vs Fabric workspace region alignment

---

Reference

Official troubleshooting guide:
https://aka.ms/foundryfabrictroubleshooting

---

If Fabric Data Agent works standalone but fails only via Foundry, in most cases this turns out to be either service-side transient failure or identity delegation edge cases rather than role misconfiguration.

If you can share region, capacity SKU, and whether multiple agents are impacted, it might help narrow this further.

[AyusmanBasu0604]

Hi @bariscihan @first of all thanks for the response as I got non from the Microsoft Learn Q&A.

FYI..

This Foundry Agent to Fabric Data tool works fine from Old Foundry (Classic Agents) We are migrating to new Foundry as the new one has workflows and the old would soon be deprecated.

Difference:

Old Foundry (Classic Agents)

1. Region: East US

2. Model: gpt-4.1

3. Tools: Fabric Data Agent  & OpenAPI (both work in proper sequence based on instructions)

New Foundry:

1. Region: East US2

2. Same Model

3. only Fabric Data Agent Tool

4. even tried with basic one liner instructions

The new one is part of a workflow and I first experienced this mid-way in an existing workflow thread where it suddenly stopped after 4th/5th response and ever since Friday it's still throwing same error even from the Agent stand alone. In the Tace Logs I could see, the part where it actually captures the response from Fabric tool is where it's throwing error. Because the operation or task: invoke_fabric_tool (azure_fabric) was successful and showing status:Ok.

typically in an agent's response there are 3 steps:

1. Invoking took

2. Capturing the response from tool

3. Message. 

As for Fabric Logs or CU utilisation of Azure foundry projects not sure where to check (please guide) 

This really is important to me and my team as my workflow solution was 90% ready and working except for the UI integration using streamlit-python 😅

very unexpected that this abruptly stopped!