phoenix-oss/llama-stack
1
0
Fork 0
forked from phoenix-oss/llama-stack-mirror
Code Pull requests Releases Packages 2 Activity Actions

fix: tracing fixes for trace context propogation across coroutines (#1522)

Browse source 
# What does this PR do?
This PR has two fixes needed for correct trace context propagation
across asycnio boundary
Fix 1: Start using context vars to store the global trace context.
This is needed since we cannot use the same trace context across
coroutines since the state is shared. each coroutine
should have its own trace context so that each of it can start storing
its state correctly.
Fix 2: Start a new span for each new coroutines started for running
shields to keep the span tree clean


## Test Plan

### Integration tests with server
LLAMA_STACK_DISABLE_VERSION_CHECK=true llama stack run
~/.llama/distributions/together/together-run.yaml
LLAMA_STACK_CONFIG=http://localhost:8321 pytest -s --safety-shield
meta-llama/Llama-Guard-3-8B --text-model
meta-llama/Llama-3.1-8B-Instruct
server logs:
https://gist.github.com/dineshyv/51ac5d9864ed031d0d89ce77352821fe
test logs:
https://gist.github.com/dineshyv/e66acc1c4648a42f1854600609c467f3
 
### Integration tests with library client
LLAMA_STACK_CONFIG=fireworks pytest -s --safety-shield
meta-llama/Llama-Guard-3-8B --text-model
meta-llama/Llama-3.1-8B-Instruct

logs: https://gist.github.com/dineshyv/ca160696a0b167223378673fb1dcefb8

### Apps test with server:
```
LLAMA_STACK_DISABLE_VERSION_CHECK=true llama stack run ~/.llama/distributions/together/together-run.yaml
python -m examples.agents.e2e_loop_with_client_tools localhost 8321
```
server logs:
https://gist.github.com/dineshyv/1717a572d8f7c14279c36123b79c5797
app logs:
https://gist.github.com/dineshyv/44167e9f57806a0ba3b710c32aec02f8
This commit is contained in:
Dinesh Yeduguru 2025-03-11 07:12:48 -07:00 • committed by GitHub
parent e3edca7739
commit ead9397e22
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
3 changed files with 54 additions and 35 deletions
Download patch file Download diff file Expand all files Collapse all files

10
llama_stack/providers/inline/agents/meta_reference/agent_instance.py
View file

@ -181,7 +181,7 @@ class ChatAgent(ShieldRunnerMixin):
return messages
async def create_and_execute_turn(self, request: AgentTurnCreateRequest) -> AsyncGenerator:
with tracing.span("create_and_execute_turn") as span:
async with tracing.span("create_and_execute_turn") as span:
span.set_attribute("session_id", request.session_id)
span.set_attribute("agent_id", self.agent_id)
span.set_attribute("request", request.model_dump_json())
@ -191,7 +191,7 @@ class ChatAgent(ShieldRunnerMixin):
yield chunk
async def resume_turn(self, request: AgentTurnResumeRequest) -> AsyncGenerator:
with tracing.span("resume_turn") as span:
async with tracing.span("resume_turn") as span:
span.set_attribute("agent_id", self.agent_id)
span.set_attribute("session_id", request.session_id)
span.set_attribute("turn_id", request.turn_id)
@ -390,7 +390,7 @@ class ChatAgent(ShieldRunnerMixin):
shields: List[str],
touchpoint: str,
) -> AsyncGenerator:
with tracing.span("run_shields") as span:
async with tracing.span("run_shields") as span:
span.set_attribute("input", [m.model_dump_json() for m in messages])
if len(shields) == 0:
span.set_attribute("output", "no shields")
@ -508,7 +508,7 @@ class ChatAgent(ShieldRunnerMixin):
content = ""
stop_reason = None
with tracing.span("inference") as span:
async with tracing.span("inference") as span:
async for chunk in await self.inference_api.chat_completion(
self.agent_config.model,
input_messages,
@ -685,7 +685,7 @@ class ChatAgent(ShieldRunnerMixin):
tool_name = tool_call.tool_name
if isinstance(tool_name, BuiltinTool):
tool_name = tool_name.value
with tracing.span(
async with tracing.span(
"tool_execution",
{
"tool_name": tool_name,