traceAI-a2a
OpenTelemetry instrumentation for the Google Agent-to-Agent (A2A) Protocol — distributed tracing across multi-agent boundaries.
The Problem
When an orchestrator agent calls a remote specialist agent over A2A, the trace breaks. The child agent starts a fresh trace. There is no parent-child link, no distributed context propagation, no visibility into what the remote agent did.
traceai-a2a fixes this.
What Gets Captured
| Data | Attribute |
|---|---|
| Task ID | gen_ai.a2a.task.id |
| Task State | gen_ai.a2a.task.state |
| Remote Agent URL | gen_ai.a2a.agent.url |
| AgentCard Name | gen_ai.a2a.agent.card.name |
| AgentCard Version | gen_ai.a2a.agent.card.version |
| Message Role | gen_ai.a2a.message.role |
| Message Parts Count | gen_ai.a2a.message.parts.count |
| SSE Streaming | gen_ai.a2a.streaming |
| Artifact Type | gen_ai.a2a.artifact.type |
| Propagated Trace ID | gen_ai.a2a.propagated_trace_id |
Install
pip install traceAI-a2a
For A2A SDK support:
pip install "traceAI-a2a[a2a]"
Quickstart
import os
from fi_instrumentation import register
from fi_instrumentation.fi_types import ProjectType
from traceai_a2a import A2AInstrumentor
os.environ["FI_API_KEY"] = "<your-api-key>"
os.environ["FI_SECRET_KEY"] = "<your-secret-key>"
# Register tracer provider
trace_provider = register(
project_type=ProjectType.OBSERVE,
project_name="my_multi_agent_app"
)
# Instrument A2A — tracing propagates across agent boundaries automatically
A2AInstrumentor().instrument(tracer_provider=trace_provider)
# Now use A2AClient as normal — both agents appear in the same trace!
from a2a.client import A2AClient
import httpx
async with httpx.AsyncClient() as http_client:
client = A2AClient(httpx_client=http_client, url="http://specialist-agent:8080")
task_id = await client.send_task(
payload={
"message": {
"role": "user",
"parts": [{"type": "text", "text": "Summarize the Q3 report"}]
}
}
)
How It Works
Orchestrator Agent Specialist Agent
───────────────── ────────────────
[span: A2A_CLIENT] [span: A2A_SERVER] ← child of A2A_CLIENT span
task_id: abc123 ────────────→ task_id: abc123
agent_url: http://... traceparent message_role: user
streaming: false header injected [span: LLM]
task_state: completed ←──────────── model: gpt-4
Both spans share the same trace ID — giving you end-to-end visibility across agent boundaries in a single FutureAGI trace view.
ASGI Server Middleware
For the receiving agent (Starlette/FastAPI):
from starlette.applications import Starlette
from traceai_a2a import A2ATracingMiddleware
app = Starlette()
app.add_middleware(A2ATracingMiddleware, tracer_provider=trace_provider)
The middleware automatically extracts the incoming traceparent header and creates child spans in the distributed trace.
See Also
Metadata
Release files for traceAI-a2a 0.1.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| traceai_a2a-0.1.0.tar.gz | 13.1 kB | With provenance | Details for traceai_a2a-0.1.0.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| traceai_a2a-0.1.0-py3-none-any.whl | Python 3 | none | any | Details for traceai_a2a-0.1.0-py3-none-any.whl |
Total release size: 28.3 kB
Release files / traceai_a2a-0.1.0.tar.gz
| Download URL | traceai_a2a-0.1.0.tar.gz |
|---|---|
| Size | 13.1 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
98498edc86c53d18d065f3f0294a3fd497de2678cdb47aae80a0a5a710fa98bc
|
|
BLAKE2b-256 checksum How to use checksums |
a1447a6c602dfc615b2d524b9ba1f032fbe598ae4fc3c9bf56c56395b4438948
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.16
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Oct 31, 2026.
Transparency logRelease files / traceai_a2a-0.1.0-py3-none-any.whl
| Download URL | traceai_a2a-0.1.0-py3-none-any.whl |
|---|---|
| Size | 15.2 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
87f4a9041d9799f8a2a95aa4b9050b25c228be08f039f4c393d6c02ca9b042af
|
|
BLAKE2b-256 checksum How to use checksums |
3f44c53368477ca3e20072fbd7c29e66a373bbc5223e1d2757a9f0da0a3a31b9
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.16
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Oct 31, 2026.
Transparency logSecurity for traceAI-a2a 0.1.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
Full provenance
All 2 files in this release have been uploaded with provenance metadata.
Attestation for traceai_a2a-0.1.0.tar.gz
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Oct 31, 2026.
Transparency logA note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.