Record streaming chunk timing metrics without mutating span events.
Updates the time_to_first_chunk histogram on the first token and
time_per_output_chunk on subsequent tokens. No span events or
attributes are written here ā the final response is recorded in
:meth:on_llm_end.
The generation chunk, when provided by LangChain.
Unique identifier for this LLM run.
Run ID of the parent chain, if any.
Additional LangChain callback keyword arguments.
