Microsoft Agent Framework

Trace Microsoft Agent Framework agents with OpenInference and send spans to Arize AX for LLM observability.

Microsoft Agent Framework is Microsoft’s open-source SDK for building production AI agents — chat, tools, multi-agent orchestration. The framework emits raw OpenTelemetry spans using GenAI semantic conventions; the openinference-instrumentation-agent-framework span processor reshapes them into the OpenInference format Arize AX expects.

This example is pinned to agent-framework-openai==1.14.4. Microsoft Agent Framework evolves quickly; if a newer release changes the API, fix the version with pip install "agent-framework-openai==1.14.4".

Prerequisites

• Python 3.10+
• An Arize AX account (sign up)
• An OPENAI_API_KEY from the OpenAI Platform

Launch Arize AX

1. Sign in to your Arize AX account.
2. From Space Settings, copy your Space ID and API Key. You will set them as ARIZE_SPACE_ID and ARIZE_API_KEY below.

Install

pip install arize-otel \
  "openinference-instrumentation-agent-framework>=0.1.13" \
  "agent-framework-openai==1.14.4"

Configure credentials

export ARIZE_SPACE_ID="<your-space-id>"
export ARIZE_API_KEY="<your-api-key>"
export ARIZE_PROJECT_NAME="microsoft-agent-framework-tracing-example"
export OPENAI_API_KEY="<your-openai-api-key>"

Setup tracing

Instrument Agent Framework with the openinference-instrumentation-agent-framework processor, regardless of model provider (OpenAI, Anthropic, Azure). Agent Framework’s own instrumentation is on by default; enable_sensitive_telemetry() adds prompts, responses, and tool arguments to the spans. Do not use a provider instrumentor such as AnthropicInstrumentor or OpenAIInstrumentor — Agent Framework drives the model through its own client layer and emits its own OpenTelemetry spans, so a provider instrumentor captures no traces. The example below uses the OpenAI client, but the same setup applies to agent-framework-anthropic and every other connector.

# instrumentation.py
import os

from agent_framework.observability import enable_sensitive_telemetry
from arize.otel import (
    BatchSpanProcessor,
    PROJECT_NAME,
    Resource,
)
from openinference.instrumentation.agent_framework import (
    AgentFrameworkToOpenInferenceProcessor,
)
from opentelemetry import trace as otel_trace
from opentelemetry.sdk.trace import TracerProvider

# Microsoft Agent Framework emits raw OpenTelemetry spans (GenAI semantic
# conventions). The reshape processor must run before the OTLP exporter,
# so build the TracerProvider manually rather than using arize.otel.register().
resource = Resource.create({
    PROJECT_NAME: os.environ["ARIZE_PROJECT_NAME"],
})
tracer_provider = TracerProvider(resource=resource)

# Reshape raw Agent Framework spans into the OpenInference format.
tracer_provider.add_span_processor(
    AgentFrameworkToOpenInferenceProcessor()
)
# Then export the reshaped spans to Arize AX.
tracer_provider.add_span_processor(
    BatchSpanProcessor(
        space_id=os.environ["ARIZE_SPACE_ID"],
        api_key=os.environ["ARIZE_API_KEY"],
    )
)

otel_trace.set_tracer_provider(tracer_provider)

# Capture prompts, responses, and tool arguments in spans. Required for span
# inputs and outputs; only enable when you trust the trace destination.
enable_sensitive_telemetry()
print("Arize AX tracing initialized for Microsoft Agent Framework.")

Run Microsoft Agent Framework

# example.py

# Importing instrumentation first ensures tracing is set up
# before `agent_framework` is imported.
from instrumentation import tracer_provider

import asyncio

from agent_framework.openai import OpenAIChatClient


async def main() -> None:
    # OpenAIChatClient reads OPENAI_API_KEY from the environment.
    agent = OpenAIChatClient(model="gpt-5.5").as_agent(
        instructions="You are a concise factual assistant.",
    )
    response = await agent.run(
        "Why is the ocean salty? Answer in two sentences."
    )
    print(response.text)


asyncio.run(main())

Expected output

Arize AX tracing initialized for Microsoft Agent Framework.
The ocean is salty because rivers continuously dissolve mineral salts from rocks and soil and carry them to the sea, where they accumulate over millions of years. Water leaves the ocean through evaporation but the salts remain, steadily concentrating until reaching today's roughly 3.5% salinity.

Verify in Arize AX

1. Open your Arize AX space and select project microsoft-agent-framework-tracing-example.
2. You should see a new trace within ~30 seconds containing an invoke_agent agent span (emitted by Agent Framework, reshaped by the OpenInference processor) wrapping a chat LLM child span. The agent span’s input is the question and its output is the answer.
3. Open the LLM span. Its input messages start with the system prompt (You are a concise factual assistant.) followed by the user question, and the span has the response and token usage attached.
4. If no traces appear, see Troubleshooting.

Check from the skill, CLI, or SDK

Confirm spans are actually reaching your Arize AX project. Use whichever fits your workflow — the skill and CLI work for any framework; the SDK check is shown for each language.

Arize skill (agent)

Install the Arize Skills plugin and let your coding agent check for you:

npx skills add Arize-ai/arize-skills

Then prompt your agent:

Use the arize-trace skill to export and analyze recent traces from my project. Confirm spans are arriving, and summarize any errors or latency issues.

AX CLI

Export recent spans for your project — any rows mean traces are landing:

ax spans export "$ARIZE_PROJECT_NAME" --space "$ARIZE_SPACE_ID" \
  --limit 5 --stdout | jq 'length'

A non-zero count confirms spans reached Arize AX. Run ax auth login first if you have not authenticated. See the ax spans reference.

SDK

Query the project’s spans and check that at least one came back.

Python

import os
from arize import ArizeClient

client = ArizeClient(api_key=os.environ["ARIZE_API_KEY"])
resp = client.spans.list(
    project=os.environ["ARIZE_PROJECT_NAME"],
    space=os.environ["ARIZE_SPACE_ID"],
    limit=5,
)
count = len(resp.spans)
print(
    f"{count} span(s) found" if count else "No spans yet — recheck setup"
)

SDK span references: Python · TypeScript · Go.

Troubleshooting

• No traces in Arize AX. Confirm ARIZE_SPACE_ID and ARIZE_API_KEY are set in the same shell that runs example.py. Enable OpenTelemetry debug logs with export OTEL_LOG_LEVEL=debug and re-run.
• Code ran but no spans appear. Import instrumentation before anything from agent_framework, so the tracer provider is set before Agent Framework creates spans. Also check that ENABLE_INSTRUMENTATION isn’t set to false in your environment.
• Spans missing inputs, outputs, or message content. Call enable_sensitive_telemetry() from agent_framework.observability. Without it, Agent Framework strips prompts, responses, and tool arguments from the spans.
• 401 from OpenAI. Verify OPENAI_API_KEY is set and has access to gpt-5.5. Swap for a model your key can call.
• API breaking changes. Microsoft Agent Framework evolves quickly; pin agent-framework-openai==1.14.4 if a newer release breaks the example.
• pip install fails with resolution-too-deep. You installed the agent-framework meta-package, which drags in every provider connector via agent-framework-core[all]. Install just the connector you need — agent-framework-openai for this example.
• Other LLM providers. Install that provider’s connector package and swap the chat client — agent-framework-anthropic, agent-framework-gemini, agent-framework-bedrock, agent-framework-mistral, agent-framework-ollama, or agent-framework-foundry (FoundryChatClient) for Azure AI Foundry. The tracing setup above is unchanged for every one of them; see the framework docs for each client’s constructor and environment variables.

Resources

Microsoft Agent Framework Documentation

OpenInference Agent Framework Span Processor

Microsoft Agent Framework GitHub