feat(batch-agent): integrate Langfuse tracing

- tracing.py: init/shutdown, trace_span, get_langfuse_callback, prompt mgmt
- main.py: init_langfuse at startup, shutdown on teardown
- redis_consumer.py: trace_span around journey_start/message/agent_trigger
- agent_runner.py: thread langfuse_handler through classify + processing LLM
- journey.py: thread langfuse_handler through _call_llm_with_tools
- llm.py: accept callbacks param, forward to LLM constructors
- requirements.txt: add langfuse>=3.0.0
This commit is contained in:
Roberto Musso
2026-03-23 08:43:15 +01:00
parent 333bba6fdd
commit 971f1dd84f
7 changed files with 347 additions and 14 deletions

View File

@@ -17,6 +17,7 @@ from typing import Any
from shared.redis import redis_client, batch_request_channel, ws_out_channel
import app.tracing as tracing
from app.ws_context import set_current_user, clear_current_user
logger = logging.getLogger(__name__)
@@ -32,15 +33,27 @@ async def _handle_journey_start(user_id: str, data: dict[str, Any]) -> None:
"""Handle a journey_start request from WS Gateway."""
from app.journey import handle_journey_start
session_id = data.get("session_id", "")
set_current_user(user_id)
try:
reply = await handle_journey_start(user_id, data)
await _publish_to_user(user_id, reply)
with tracing.trace_span(
name="journey_start",
user_id=user_id,
session_id=session_id,
input=data.get("directory", ""),
metadata={"data_types": data.get("data_types", [])},
tags=["journey"],
) as span:
langfuse_handler = tracing.get_langfuse_callback()
reply = await handle_journey_start(user_id, data, langfuse_handler=langfuse_handler)
span.update(output=reply.get("message", "")[:500])
await _publish_to_user(user_id, reply)
tracing.flush()
except Exception as exc:
logger.error("batch-agent: journey_start failed user=%s: %s", user_id, exc)
await _publish_to_user(user_id, {
"type": "journey_reply",
"session_id": data.get("session_id", ""),
"session_id": session_id,
"message": f"Journey setup failed: {exc}",
"done": True,
"prompt_template": None,
@@ -53,15 +66,26 @@ async def _handle_journey_message(user_id: str, data: dict[str, Any]) -> None:
"""Handle a journey_message from WS Gateway."""
from app.journey import handle_journey_message
session_id = data.get("session_id", "")
set_current_user(user_id)
try:
reply = await handle_journey_message(user_id, data)
await _publish_to_user(user_id, reply)
with tracing.trace_span(
name="journey_message",
user_id=user_id,
session_id=session_id,
input=data.get("message", "")[:200],
tags=["journey"],
) as span:
langfuse_handler = tracing.get_langfuse_callback()
reply = await handle_journey_message(user_id, data, langfuse_handler=langfuse_handler)
span.update(output=reply.get("message", "")[:500])
await _publish_to_user(user_id, reply)
tracing.flush()
except Exception as exc:
logger.error("batch-agent: journey_message failed user=%s: %s", user_id, exc)
await _publish_to_user(user_id, {
"type": "journey_reply",
"session_id": data.get("session_id", ""),
"session_id": session_id,
"message": f"Journey processing failed: {exc}",
"done": True,
"prompt_template": None,
@@ -74,15 +98,28 @@ async def _handle_agent_trigger(user_id: str, data: dict[str, Any]) -> None:
"""Handle an agent_trigger request from the REST route (forwarded via Redis)."""
from app.agent_runner import run_local_agent
run_context = data.get("run_context", {})
agent_id = run_context.get("agent_id", "")
set_current_user(user_id)
try:
await run_local_agent(user_id, data)
with tracing.trace_span(
name="agent_trigger",
user_id=user_id,
trace_id=run_context.get("run_id"),
input={"agent_id": agent_id, "directory": data.get("directory", "")},
metadata={"data_types": data.get("data_types", [])},
tags=["batch", "agent_run"],
) as span:
langfuse_handler = tracing.get_langfuse_callback()
await run_local_agent(user_id, data, langfuse_handler=langfuse_handler)
span.update(output={"status": "completed"})
tracing.flush()
except Exception as exc:
logger.error("batch-agent: agent_trigger failed user=%s: %s", user_id, exc)
await _publish_to_user(user_id, {
"type": "run_complete",
"status": "error",
"run_context": data.get("run_context", {}),
"run_context": run_context,
})
finally:
clear_current_user()