mirror of
https://github.com/BerriAI/litellm.git
synced 2025-04-26 19:24:27 +00:00
Improve rpm check on keys (#8301)
* fix(parallel_request_limiter.py): initial commit that solves the rpm limit check on keys Fixes https://github.com/BerriAI/litellm/issues/6938 * fix(parallel_request_limiter.py): simpler approach - just increment RPM in pre call hook instead of on success * fix(parallel_request_limiter.py): pass testing * fix: fix linting error * fix(parallel_request_limiter.py): fix parallel request check for keys
This commit is contained in:
parent
f66bcce2fa
commit
d08536982e
5 changed files with 92 additions and 140 deletions
|
@ -414,7 +414,9 @@ def bytes_to_mb(bytes_value: int):
|
|||
|
||||
|
||||
# helpers used by parallel request limiter to handle model rpm/tpm limits for a given api key
|
||||
def get_key_model_rpm_limit(user_api_key_dict: UserAPIKeyAuth) -> Optional[dict]:
|
||||
def get_key_model_rpm_limit(
|
||||
user_api_key_dict: UserAPIKeyAuth,
|
||||
) -> Optional[Dict[str, int]]:
|
||||
if user_api_key_dict.metadata:
|
||||
if "model_rpm_limit" in user_api_key_dict.metadata:
|
||||
return user_api_key_dict.metadata["model_rpm_limit"]
|
||||
|
@ -428,7 +430,9 @@ def get_key_model_rpm_limit(user_api_key_dict: UserAPIKeyAuth) -> Optional[dict]
|
|||
return None
|
||||
|
||||
|
||||
def get_key_model_tpm_limit(user_api_key_dict: UserAPIKeyAuth) -> Optional[dict]:
|
||||
def get_key_model_tpm_limit(
|
||||
user_api_key_dict: UserAPIKeyAuth,
|
||||
) -> Optional[Dict[str, int]]:
|
||||
if user_api_key_dict.metadata:
|
||||
if "model_tpm_limit" in user_api_key_dict.metadata:
|
||||
return user_api_key_dict.metadata["model_tpm_limit"]
|
||||
|
|
Loading…
Add table
Add a link
Reference in a new issue