scriptoriumadmin/litellm
1
0
Fork 0
mirror of https://github.com/BerriAI/litellm.git synced 2026-10-10 03:28:53 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 86

fix(azure): normalize equivalent default retry cache entries

Browse source 
Signed-off-by: apex-mochen <2756823972@qq.com>
This commit is contained in:
apex-mochen 2026-10-03 09:21:10 +08:00
parent 4b9454dd1a
commit 82d13fbe6b
2 changed files with 9 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

5
litellm/llms/azure/common_utils.py
View file

@ -491,7 +491,10 @@ class BaseAzureLLM(BaseOpenAILLM):
client_initialization_params: Final[dict] = locals()
client_initialization_params["is_async"] = _is_async
_lp: Final = litellm_params or {}
client_initialization_params["max_retries"] = _lp.get("max_retries")
configured_max_retries: Final = _lp.get("max_retries")
client_initialization_params["max_retries"] = (
DEFAULT_MAX_RETRIES if configured_max_retries is None else configured_max_retries
)
_ad_provider: Final = _lp.get("azure_ad_token_provider")
_ad_token: Final = _lp.get("azure_ad_token")
_client_secret: Final = _lp.get("client_secret")

5
tests/unit/llms/azure/test_azure_common_utils.py
View file

@ -2235,6 +2235,11 @@ async def test_azure_client_cache_respects_retry_limit(
assert initial.max_retries == (
DEFAULT_MAX_RETRIES if first_retries is None else first_retries
)
if first_retries is None:
assert llm.get_azure_openai_client(**kwargs, litellm_params={}) is initial
assert llm.get_azure_openai_client(
**kwargs, litellm_params={"max_retries": DEFAULT_MAX_RETRIES}
) is initial
with pytest.raises(RateLimitError):
response = no_retry.chat.completions.create(
model="test-deployment", messages=[{"role": "user", "content": "hi"}]