Agent Framework RC error SpanAttributes.LLM_SYSTEM, #4276
Description
Description
Hey team has anyone hit this using agent-framework==1.0.0rc1?
pyproject.toml
#pyproject.toml
[project]
name = "bug"
version = "0.0.1"
description = "agent-framework==1.0.0rc1 + semconv_ai==0.4.14"
requires-python = ">=3.13"
dependencies = [
"agent-framework==1.0.0rc1"]
[tool.uv]
prerelease = "allow"
Agent
"""
agent.py — Intenta crear un agente dummy con agent-framework==1.0.0rc1.
run agent:
uv run python agent.py
Variables de entorno requeridas:
AZURE_AI_PROJECT_ENDPOINT — e.g. https://<hub>.services.ai.azure.com/...
AZURE_AI_MODEL — nombre del deployment, e.g. gpt-4o
"""
import asyncio
import os
from azure.identity import DefaultAzureCredential
from agent_framework.azure import AzureAIClient
ENDPOINT = os.environ["AZURE_AI_PROJECT_ENDPOINT"]
MODEL = os.environ.get("AZURE_AI_MODEL", "gpt-4o")
async def main():
client = AzureAIClient(
project_endpoint=ENDPOINT,
model_deployment_name=MODEL,
credential=DefaultAzureCredential(),
)
agent = client.as_agent(
name="dummy",
instructions="""
1) A robot may not injure a human being...
2) A robot must obey orders given it by human beings...
3) A robot must protect its own existence...
Give me the TLDR in exactly 5 words.
""",
)
result = await agent.run("Summarize the Three Laws of Robotics")
print(result)
asyncio.run(main())The error happens at import time before any Agent or AzureAIClient is even constructed.
Agent-framework pulls in opentelemetry-semantic-conventions-ai as a transitive dependency, and the version it resolves to (0.4.x) no longer has LLM_SYSTEM and related attributes that the framework references at the top level of observability.
Has anyone else run into this? Is there a pinned version of semconv_ai that works, or a workaround you're using or is just me ?
Code Sample
"""
agent.py — Intenta crear un agente dummy con agent-framework==1.0.0rc1.
run agent:
uv run python agent.py
Variables de entorno requeridas:
AZURE_AI_PROJECT_ENDPOINT — e.g. https://<hub>.services.ai.azure.com/...
AZURE_AI_MODEL — nombre del deployment, e.g. gpt-4o
"""
import asyncio
import os
from azure.identity import DefaultAzureCredential
from agent_framework.azure import AzureAIClient
ENDPOINT = os.environ["AZURE_AI_PROJECT_ENDPOINT"]
MODEL = os.environ.get("AZURE_AI_MODEL", "gpt-4o")
async def main():
client = AzureAIClient(
project_endpoint=ENDPOINT,
model_deployment_name=MODEL,
credential=DefaultAzureCredential(),
)
agent = client.as_agent(
name="dummy",
instructions="""
1) A robot may not injure a human being...
2) A robot must obey orders given it by human beings...
3) A robot must protect its own existence...
Give me the TLDR in exactly 5 words.
""",
)
result = await agent.run("Summarize the Three Laws of Robotics")
print(result)
asyncio.run(main())
Error Messages / Stack Traces
"system_name": (SpanAttributes.LLM_SYSTEM, None, False, None),
^^^^^^^^^^^^^^^^^^^^^^^^^
AttributeError: type object 'SpanAttributes' has no attribute 'LLM_SYSTEM'Package Versions
gent-framework==1.0.0rc1
Python Version
Python 3.13.11
Additional Context
No response