[FEATURE]: Propagate ml_app with LLMObs.activate_distributed_headers

[shotarok]

Package Name

llmobs

Package Version(s)

2.17.2

Describe the goal of the feature

ml_app is automatically propagated from a parent LLMObs span to a child LLMObs span in a single process. However, ml_app is not propagated at distributed tracing with LLMObs.inject_distributed_headers and LLMObs.activate_distributed_headers. The goal is to propagate ml_app automatically from a parent span to a child span, even with the distributed headers.

Is your feature request related to a problem?

No response

Describe alternatives you've considered

Right now, I have defined a custom header to propagate ml_app and reuse it when creating a child span. It isn't easy to ensure the same pattern in the code base.

Additional context

No response

[Yun-Kim]

Hi @shotarok, thanks for reaching out! We are actually working on this feature right now so hopefully we'll have this be available in the next few weeks.