temporalio / temporalio/sdk-python

[Feature Request] Support for custom tracing providers

Open
#1,130 1 comment 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

enhancement
Dominant language
Python
Stars
1.2k
Forks
241
Avg merge
3d 21h
Merged PRs (30d)
55

Description

I'd like to add a custom tracing provider for the OpenAI Agents SDK. Importantly, this tracing provider is supported by the OpenAI Agents SDK itself.

The solution I'd like

I'd like to be able to pass the tracing provider to the OpenAI Agents Plugin when initializing the client e.g.

from temporalio.client import Client
from temporalio.contrib.openai_agents import OpenAIAgentsPlugin
from custom_tracing_provider import CustomTraceProvider

plugin = OpenAIAgentsPlugin(
     trace_provider=CustomTraceProvider()
)


client_config = {
    "namespace": NAMESPACE,
    "plugins": [plugin],
    "API_KEY": os.environ.get("TEMPORAL_API_KEY")
}

_client_instance = await Client.connect(ADDRESS, **client_config)

Given the temporal tracing provider today is implemented on top of the DefaultTracingProvider class provided by OpenAI as shown below, perhaps it wouldn't be too hard to just create a factory that takes any supported tracing provider and returns a Temporal supported one.


class TemporalTraceProvider(DefaultTraceProvider):
    """A trace provider that integrates with Temporal workflows."""

    def __init__(self, auto_close_in_workflows: bool = False):
        """Initialize the TemporalTraceProvider."""
        super().__init__()
        self._original_provider = cast(DefaultTraceProvider, get_trace_provider())
        self._multi_processor = _TemporalTracingProcessor(
            self._original_provider._multi_processor,
            auto_close_in_workflows,
        )

    def time_iso(self) -> str:
        """Return the current deterministic time in ISO 8601 format."""
        if workflow.in_workflow():
            return workflow.now().isoformat()
        return super().time_iso()

    def gen_trace_id(self) -> str:
        """Generate a new trace ID."""
        if workflow.in_workflow():
            try:
                """Generate a new trace ID."""
                return f"trace_{_workflow_uuid()}"
            except ReadOnlyContextError:
                return f"trace_{uuid.uuid4().hex}"
        return super().gen_trace_id()

    def gen_span_id(self) -> str:
        """Generate a span ID."""
        if workflow.in_workflow():
            try:
                """Generate a deterministic span ID."""
                return f"span_{_workflow_uuid()}"
            except ReadOnlyContextError:
                return f"span_{uuid.uuid4().hex[:24]}"
        return super().gen_span_id()

    def gen_group_id(self) -> str:
        """Generate a group ID."""
        if workflow.in_workflow():
            try:
                """Generate a deterministic group ID."""
                return f"group_{_workflow_uuid()}"
            except ReadOnlyContextError:
                return f"group_{uuid.uuid4().hex[:24]}"
        return super().gen_group_id()

    def __enter__(self):
        """Enter the context of the Temporal trace provider."""
        return self

    def __exit__(
        self,
        exc_type: type[BaseException],
        exc_val: BaseException,
        exc_tb: TracebackType,
    ):
        """Exit the context of the Temporal trace provider."""
        self._multi_processor.shutdown()

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by reading OpenAIAgentsPlugin initialization and the shown TemporalTraceProvider implementation, then determine how a supplied tracing provider can be integrated while preserving Temporal workflow behavior. Done means callers can pass a supported custom provider when creating the plugin and tracing continues to work correctly in workflows.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
observability-sre
Issue type
Feature
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
38/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.