scriptoriumadmin/litellm
1
0
Fork 0
mirror of https://github.com/BerriAI/litellm.git synced 2026-10-30 06:43:27 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 249

fix(proxy): tolerate a StandardLoggingPayload without total_tokens

Browse source 
Some call types (e.g. aretrieve_batch) build a payload without total_tokens,
so indexing it raised KeyError and tripped the cost-tracking error path. Read
it with a .get default of 0 like the DB-side write, suppressing just the one
overload-unknown diagnostic with a reason.
This commit is contained in:
Srivatsa03 2026-10-11 22:57:34 -05:00
parent f949f1735b
commit fe6edbef4c
1 changed files with 3 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

4
litellm/proxy/hooks/proxy_track_cost_callback.py
View file

@ -741,7 +741,9 @@ def _extract_total_tokens(sl_object: StandardLoggingPayload | None) -> int:
"""
if sl_object is None:
return 0
return sl_object["total_tokens"]
return sl_object.get( # pyright: ignore[reportUnknownMemberType] # StandardLoggingPayload.get overload is partially unknown; total_tokens is a required int
"total_tokens", 0
)
async def _update_database_and_spend_counters(