Description
Calling AIProjectClient.agents.create_version() to deploy a hosted agent consistently returns an HTTP 500 server_error with no actionable details. The error persists across retries and different SDK configurations. The container image is valid, pushed to ACR, and the agent previously deployed successfully on the older SDK stack.
Code Sample
from azure.identity import DefaultAzureCredential
from azure.ai.projects import AIProjectClient
from azure.ai.projects.models import (
HostedAgentDefinition,
ProtocolVersionRecord,
AgentProtocol,
)
client = AIProjectClient(
endpoint="<FOUNDRY_PROJECT_ENDPOINT>",
credential=DefaultAzureCredential(),
allow_preview=True,
)
agent = client.agents.create_version(
agent_name="agent-name",
definition=HostedAgentDefinition(
container_protocol_versions=[
ProtocolVersionRecord(
protocol=AgentProtocol.RESPONSES,
version="1.0.0",
)
],
cpu="1",
memory="2Gi",
image="<ACR>.azurecr.io/<IMAGE>:<TAG>",
environment_variables={...}, # 16 env vars
tools=[
{
"type": "mcp",
"server_label": "WorkIQUser",
"project_connection_id": "<MCP_CONNECTION_ID>",
}
],
),
)
Error Messages / Stack Traces
HttpResponseError: (server_error) An error occurred while processing your request.
You can retry your request, or contact us if the error persists.
Code: server_error
Package Versions
SDK: azure-ai-projects==2.1.0, agent-framework==1.2.2, agent-framework-foundry-hosting==1.0.0a260429
Python Version
Python 3.12
Additional Context
No response
Description
Calling
AIProjectClient.agents.create_version()to deploy a hosted agent consistently returns an HTTP 500server_errorwith no actionable details. The error persists across retries and different SDK configurations. The container image is valid, pushed to ACR, and the agent previously deployed successfully on the older SDK stack.Code Sample
from azure.identity import DefaultAzureCredential from azure.ai.projects import AIProjectClient from azure.ai.projects.models import ( HostedAgentDefinition, ProtocolVersionRecord, AgentProtocol, ) client = AIProjectClient( endpoint="<FOUNDRY_PROJECT_ENDPOINT>", credential=DefaultAzureCredential(), allow_preview=True, ) agent = client.agents.create_version( agent_name="agent-name", definition=HostedAgentDefinition( container_protocol_versions=[ ProtocolVersionRecord( protocol=AgentProtocol.RESPONSES, version="1.0.0", ) ], cpu="1", memory="2Gi", image="<ACR>.azurecr.io/<IMAGE>:<TAG>", environment_variables={...}, # 16 env vars tools=[ { "type": "mcp", "server_label": "WorkIQUser", "project_connection_id": "<MCP_CONNECTION_ID>", } ], ), )Error Messages / Stack Traces
Package Versions
SDK:
azure-ai-projects==2.1.0,agent-framework==1.2.2,agent-framework-foundry-hosting==1.0.0a260429Python Version
Python 3.12
Additional Context
No response