The LangchainCallbackHandler supports statistical delay of the first token

[Hacky-DH]

Describe the feature or potential improvement

In the on_llm_new_token method of the LangchainCallbackHandler class, update the completion_start_time parameter of the generation using the current time. This way, you will be able to see the value of Time to First Token on the front-end page.

Additional information

No response

[hassiebp]

Thanks for raising this, indeed a good improvement!

[marcklingen]

tracking this here: #3432

[hassiebp]

@Hacky-DH Thanks again for your report, this feature has been now released 👍🏾 https://github.com/langfuse/langfuse-python/releases/tag/v2.51.0