Adrian Cole

An anecdote I noticed is [kubeday japan](https://kubedayjapan2024.sched.com/) currently has no session on otel, nor an obvious event around the timing, even unofficial. This is despite some really cool leads like...

thanks for the help, @jack-berg. the advice was spot-on and appreciate your opening an issue on the gap!

@mikeldking I didn't know whether to call this a bug or a feat, so apologies on being a mix. you can feel free to correct the title and labels

Thanks for all the help @GALLLASMILAN and good luck with whatever's next! @pilartomas do you have an opinion about switching span modeling from the current (events as spans, span name...

Mostly there, but seems still some "start/stop" spans which are not idiomatic, yet. Here are the requirements I am using now, which are just version bumps from before: ``` beeai-framework~=0.1.40...

nice, thanks! ``` beeai-framework~=0.1.40 litellm~=1.76.1 httpx~=0.28.1 elastic-opentelemetry~=1.7.0 openinference-instrumentation-beeai~=0.1.10 opentelemetry-instrumentation-httpx~=0.57b0 ``` fyi: since beeai has a direct dep on litellm, it now brings in a new platform dep if you update...

@Tomas2D yeah it seems at the moment there's a direct coupling to litellm for all main model providers. https://github.com/i-am-bee/beeai-framework/blob/89c4951665f36bac52d9c5493e79737c034f8548/python/beeai_framework/adapters/openai/backend/chat.py#L17 The dependency chain is: beeai-framework (~0.1.40) └── litellm (=1.69.3) └── fastuuid...

yeah you probably would need an end user complaint or few to justify having a different impl under your abstraction or rather litellm wrapper than this build glitch. thanks for...

yes it is still relevant, and we need to get to a "do it" or "don't do it" I think, because the topic space is really binary in nature. if...

I have read the CLA Document and I hereby sign the CLA