Getting this error: server had an error processing your request

Question

Getting this error: server had an error processing your request

Sweekriti Pant 0 Microsoft Employee

Error

There was an error generating a response. Chat history can't be saved at this time. Error code: 500 - {'error': {'message': 'The server had an error processing your request. Sorry about that! You can retry your request, or contact us through an Azure support request at: https://go.microsoft.com/fwlink/?linkid=2213926 if you keep seeing this error. (Please include the request ID 028c0761-4913-490e-aa53-f523e9ba9e38 in your email.)', 'type': 'server_error', 'param': None, 'code': 'server_error'}}

When I am asking a question to my bot, i am getting this response. How to fix this?

Karnam Venkata Rajeswari 3,840 Reputation points Microsoft External Staff Moderator

2026-06-18T12:53:27.6333333+00:00
Hello @Sweekriti Pant ,

Welcome to Microsoft Q&A .Thank you for reaching out to us.

Based on the current behavior, the issue appears to be isolated to a specific Azure AI Foundry agent. This assessment is based on the fact that another agent within the same project is operating successfully using the same GPT-5.2 deployment, the same region (East US) and the same Responses API workflow.

The error message “Chat history can’t be saved at this time” may indicate that the failure occurs during the agent execution workflow, such as while processing conversation state or while invoking an associated tool or dependency.

The following might be the probable causes for this error to be encountered-

Agent dependency or tool issue A resource used by the affected agent may be unavailable, changed, or incorrectly configured, including:

Azure AI Search indexes

Knowledge sources

Uploaded files

Vector stores

Function/tool definitions

MCP tools or external connectors

Authentication, permissions or connection settings

Agent configuration changes A recent update may have introduced an inconsistency, such as:

Changes to instructions or system prompts

Updates to tool definitions or schemas

Structured output configuration changes

Agent republishing or configuration updates

gent-specific execution or metadata issue -A problem affecting the agent definition, configuration state or execution path may be impacting only this particular agent.

Recent changes or dependency updates Since the issue started after previously working, we suggest to please review any changes made before the failure began, including agent updates, resource modifications or dependency changes.

Please check if the following steps help-

Testing the affected agent directly in Azure AI Foundry Playground

Run a simple prompt against the failing agent directly from the Foundry portal.

If the same issue reproduces, this helps confirm that the behavior is associated with the agent configuration, attached resources, or backend execution rather than only the application layer.

Performing a baseline model test

Please send a request to the same GPT-5.2 deployment using the Responses API without agent_reference.

This helps validate whether the underlying model deployment and API workflow function independently of the agent configuration.

Comparing the working and failing agents Review differences between both agents, including:

Instructions/system prompt

Tools and function definitions

Knowledge sources

Azure AI Search configuration

Uploaded files

Vector stores

Connectors/MCP tools

Structured output settings

Agent-specific runtime configuration

The difference between the two agents is likely to provide the most useful direction for further investigation.

Validating dependent resources Please confirm that all resources referenced by the failing agent are:

Available and accessible

Correctly configured

Properly authenticated

Not deleted, expired or unavailable

Simplifying and isolate the configuration

Temporarily remove tools and knowledge sources from the affected agent.

Test with a basic prompt-only configuration.

Add components back incrementally to identify whether a specific dependency contributes to the failure.

Creating a new test agent

Create a new Azure AI Foundry agent using the same GPT-5.2 deployment with minimal configuration.

Test the new agent with a simple prompt.

If the new agent works while the existing agent continues to fail, this would strongly suggest that the issue is specific to the original agent configuration, metadata, or attached resources.

Reviewing recent changes Verify whether any changes occurred before the issue started:

Agent instructions updated

Tools added or modified

Knowledge sources changed

Files or indexes updated

Connections or credentials changed

Agent republished

SDK or application dependency updates

Depending on the application scenario, the following options can be considered as workarounds-

Use the working agent temporarily.

Clone the working agent configuration and update the application reference.

Recreate or refresh any dependency identified during troubleshooting.

Recreate the affected agent if configuration corruption is suspected.

The following references might be helpful , please check them out

Microsoft Foundry documentation | Microsoft Learn

Build with agents, conversations, and responses in Foundry Agent Service - Microsoft Foundry | Microsoft Learn

Use the Azure OpenAI Responses API - Microsoft Foundry | Microsoft Learn

Please let us know if the response was helpful

Thank you
Karnam Venkata Rajeswari 3,840 Reputation points Microsoft External Staff Moderator

2026-06-18T13:08:20.14+00:00
Hello @Sweekriti Pant ,

Thanks for sharing the request ID. To help narrow down the cause of the HTTP 500 error, could you please help us with the following details:

Environment: Is the bot using Azure OpenAI or Azure AI Foundry and if this is occurring in the playground, an Agent, a Bot/Copilot or via direct API calls?

Request details: API version, API type (Chat Completions or Responses API),and model/deployment are being used?

Error scope: Does the error occur for all requests or only specific prompts/features (e.g., large inputs, attachments, tool calling)? Does a retry succeed, and was it working previously?

Deployment/region: Is the issue limited to a specific deployment or region?

Thank you
Sweekriti Pant 0 Reputation points Microsoft Employee

2026-06-18T13:16:33.2733333+00:00
@Karnam Venkata Rajeswari These are the following details:

The bot uses Azure AI Foundry (Responses API with agent_reference). It's a custom python web app deployed on Azure App Service, making direct API calls to Foundry via sdk. The agent is configured in Azure AI Foundry and called via client.responses.create() with an agent_reference in extra_body.\

API type: Responses API (client.responses.create()) , Model/deployment: gpt-5.2

All requests. Retry did not succeed. It was working previously, started seeing this error today onwards on all requests.

It is limited on just 1 foundry agent in the project, another foundry agent using the same gpt model deployment & in the same region (eastus) is working fine.

Please let me know if any more details are needed from my end.
Karnam Venkata Rajeswari 3,840 Reputation points Microsoft External Staff Moderator

2026-06-18T13:19:52.2233333+00:00

Hello @Sweekriti Pant ,

Thank you for quickly assisting us with the details. Please allow us some time while we work on this

Thank you
Karnam Venkata Rajeswari 3,840 Reputation points Microsoft External Staff Moderator

2026-06-19T14:28:34.4+00:00

Hello @Sweekriti Pant ,

Following up to know if the response was helpful

Thank you