diff --git a/litellm/proxy/auth/auth_exception_handler.py b/litellm/proxy/auth/auth_exception_handler.py index 2eb25cc6105..aa491e8dd42 100644 --- a/litellm/proxy/auth/auth_exception_handler.py +++ b/litellm/proxy/auth/auth_exception_handler.py @@ -29,6 +29,7 @@ from litellm.proxy.auth.auth_utils import ( # noqa: F401 # legacy module expor ) from litellm.proxy.auth.model_access_denied import ModelAccessDeniedHTTPException from litellm.proxy.auth.resolvers.exceptions import KeyNotFoundError +from litellm.proxy.common_utils.openai_error_payload import openai_error_param from litellm.proxy.db.exception_handler import PrismaDBExceptionHandler from litellm.types.services import ServiceTypes @@ -60,14 +61,14 @@ def _as_proxy_exception(e: Exception) -> ProxyException: message=str(e.detail), internal_message=e.internal_message, type=ProxyErrorTypes.auth_error, - param="None", + param=None, code=e.status_code, ) if isinstance(e, HTTPException): return ProxyException( message=getattr(e, "detail", f"Authentication Error({e})"), type=ProxyErrorTypes.auth_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=getattr(e, "status_code", status.HTTP_401_UNAUTHORIZED), ) if isinstance(e, ProxyException): @@ -77,7 +78,7 @@ def _as_proxy_exception(e: Exception) -> ProxyException: return ProxyException( message="Authentication Error, " + str(e), type=ProxyErrorTypes.auth_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=status.HTTP_401_UNAUTHORIZED, ) diff --git a/litellm/proxy/auth/handle_jwt.py b/litellm/proxy/auth/handle_jwt.py index 9d800a1fd63..7bd7b1226b1 100644 --- a/litellm/proxy/auth/handle_jwt.py +++ b/litellm/proxy/auth/handle_jwt.py @@ -225,7 +225,7 @@ def jwks_unavailable_exception(error: JWKSUnreachableError) -> ProxyException: f"unreachable, so the JWT signature could not be verified. Please retry shortly. Error: {error}" ), type=ProxyErrorTypes.auth_provider_unavailable, - param="None", + param=None, code=status.HTTP_503_SERVICE_UNAVAILABLE, ) diff --git a/litellm/proxy/db/exception_handler.py b/litellm/proxy/db/exception_handler.py index 5dfcaa4b2f8..da32df87444 100644 --- a/litellm/proxy/db/exception_handler.py +++ b/litellm/proxy/db/exception_handler.py @@ -435,7 +435,7 @@ class PrismaDBExceptionHandler: return ProxyException( message=PrismaDBExceptionHandler.database_unavailable_message(e), type=ProxyErrorTypes.no_db_connection, - param="None", + param=None, code=HTTPStatus.SERVICE_UNAVAILABLE.value, ) diff --git a/litellm/proxy/health_endpoints/_health_endpoints.py b/litellm/proxy/health_endpoints/_health_endpoints.py index 53c54fe15bf..1d09a56ed3a 100644 --- a/litellm/proxy/health_endpoints/_health_endpoints.py +++ b/litellm/proxy/health_endpoints/_health_endpoints.py @@ -47,6 +47,7 @@ from litellm.proxy.auth.auth_utils import ( from litellm.proxy.auth.model_checks import get_key_models from litellm.proxy.auth.user_api_key_auth import user_api_key_auth from litellm.proxy.common_utils import http_parsing_utils +from litellm.proxy.common_utils.openai_error_payload import openai_error_param from litellm.proxy.db.db_lookup_gate import db_lookup_stall_tracker from litellm.proxy.db.exception_handler import PrismaDBExceptionHandler from litellm.proxy.db.health_check_latest import ( @@ -687,7 +688,7 @@ async def health_services_endpoint( raise ProxyException( message=getattr(e, "detail", f"Authentication Error({e})"), type=ProxyErrorTypes.auth_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=getattr(e, "status_code", status.HTTP_500_INTERNAL_SERVER_ERROR), ) elif isinstance(e, ProxyException): @@ -695,7 +696,7 @@ async def health_services_endpoint( raise ProxyException( message="Authentication Error, " + str(e), type=ProxyErrorTypes.auth_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=status.HTTP_500_INTERNAL_SERVER_ERROR, ) diff --git a/litellm/proxy/management_endpoints/internal_user_endpoints.py b/litellm/proxy/management_endpoints/internal_user_endpoints.py index d6bf9220551..b8c8ccf645b 100644 --- a/litellm/proxy/management_endpoints/internal_user_endpoints.py +++ b/litellm/proxy/management_endpoints/internal_user_endpoints.py @@ -46,6 +46,7 @@ from litellm.proxy.auth.password_policy import ( ) from litellm.proxy.auth.user_api_key_auth import user_api_key_auth from litellm.proxy.common_utils.auth_cache_invalidation_pubsub import evict_and_broadcast +from litellm.proxy.common_utils.openai_error_payload import openai_error_param from litellm.proxy.common_utils.user_api_key_cache import ( AUTH_OBJECTS_TARGET, object_permission_cache_key, @@ -1733,7 +1734,7 @@ async def user_update( raise ProxyException( message=getattr(e, "detail", f"Authentication Error({e})"), type=ProxyErrorTypes.auth_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=getattr(e, "status_code", status.HTTP_400_BAD_REQUEST), ) elif isinstance(e, ProxyException): @@ -1741,7 +1742,7 @@ async def user_update( raise ProxyException( message="Authentication Error, " + str(e), type=ProxyErrorTypes.auth_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=status.HTTP_400_BAD_REQUEST, ) diff --git a/litellm/proxy/management_endpoints/key_management_endpoints.py b/litellm/proxy/management_endpoints/key_management_endpoints.py index 0ad4258fb04..055d7d1f80d 100644 --- a/litellm/proxy/management_endpoints/key_management_endpoints.py +++ b/litellm/proxy/management_endpoints/key_management_endpoints.py @@ -83,6 +83,7 @@ from litellm.proxy.common_utils.config_sync_pubsub import ( coordination_redis_cache, publish_config_change, ) +from litellm.proxy.common_utils.openai_error_payload import openai_error_param from litellm.proxy.common_utils.rbac_utils import check_org_admin_can_generate_keys from litellm.proxy.common_utils.timezone_utils import get_budget_reset_time from litellm.proxy.common_utils.user_api_key_cache import AUTH_OBJECTS_TARGET, UserApiKeyCache @@ -4090,7 +4091,7 @@ async def update_key_fn( raise ProxyException( message=getattr(e, "detail", f"Authentication Error({e})"), type=ProxyErrorTypes.auth_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=getattr(e, "status_code", status.HTTP_400_BAD_REQUEST), ) elif isinstance(e, ProxyException): @@ -4098,7 +4099,7 @@ async def update_key_fn( raise ProxyException( message="Authentication Error, " + str(e), type=ProxyErrorTypes.auth_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=status.HTTP_400_BAD_REQUEST, ) @@ -4969,7 +4970,7 @@ async def info_key_fn( raise ProxyException( message="Internal server error from /key/info. Details are in the proxy server logs.", type="internal_error", - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=status.HTTP_500_INTERNAL_SERVER_ERROR, ) @@ -7244,7 +7245,7 @@ async def list_keys( raise ProxyException( message=getattr(e, "detail", f"error({e})"), type=ProxyErrorTypes.internal_server_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=getattr(e, "status_code", fastapi.status.HTTP_500_INTERNAL_SERVER_ERROR), ) elif isinstance(e, ProxyException): @@ -7252,7 +7253,7 @@ async def list_keys( raise ProxyException( message="Authentication Error, " + str(e), type=ProxyErrorTypes.internal_server_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=fastapi.status.HTTP_500_INTERNAL_SERVER_ERROR, ) @@ -7402,7 +7403,7 @@ async def key_aliases( raise ProxyException( message=getattr(e, "detail", f"error({e})"), type=ProxyErrorTypes.internal_server_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=getattr(e, "status_code", status.HTTP_500_INTERNAL_SERVER_ERROR), ) elif isinstance(e, ProxyException): @@ -7410,7 +7411,7 @@ async def key_aliases( raise ProxyException( message="Authentication Error, " + str(e), type=ProxyErrorTypes.internal_server_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=status.HTTP_500_INTERNAL_SERVER_ERROR, ) @@ -8187,7 +8188,7 @@ async def key_health( raise ProxyException( message=f"Key health check failed: {e}", type=ProxyErrorTypes.internal_server_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=status.HTTP_500_INTERNAL_SERVER_ERROR, ) diff --git a/litellm/proxy/management_endpoints/model_management_endpoints.py b/litellm/proxy/management_endpoints/model_management_endpoints.py index e625ea6f9f2..ac1bd6ce909 100644 --- a/litellm/proxy/management_endpoints/model_management_endpoints.py +++ b/litellm/proxy/management_endpoints/model_management_endpoints.py @@ -75,6 +75,7 @@ from litellm.proxy.common_utils.encrypt_decrypt_utils import ( decrypt_value_helper, encrypt_value_helper, ) +from litellm.proxy.common_utils.openai_error_payload import openai_error_param from litellm.proxy.common_utils.user_api_key_cache import UserApiKeyCache from litellm.proxy.db.routing_prisma_wrapper import WriterPinnedClient from litellm.proxy.management.teams.authz import TEAM_ADMIN_ONLY, is_team_admin @@ -2612,7 +2613,7 @@ async def delete_model( raise ProxyException( message=getattr(e, "detail", f"Authentication Error({e})"), type=ProxyErrorTypes.auth_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=getattr(e, "status_code", status.HTTP_400_BAD_REQUEST), ) elif isinstance(e, ProxyException): @@ -2620,7 +2621,7 @@ async def delete_model( raise ProxyException( message="Authentication Error, " + str(e), type=ProxyErrorTypes.auth_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=status.HTTP_400_BAD_REQUEST, ) @@ -2875,7 +2876,7 @@ async def add_new_model( raise ProxyException( message=getattr(e, "detail", f"Authentication Error({e})"), type=ProxyErrorTypes.auth_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=getattr(e, "status_code", status.HTTP_400_BAD_REQUEST), ) elif isinstance(e, ProxyException): @@ -2883,7 +2884,7 @@ async def add_new_model( raise ProxyException( message="Authentication Error, " + str(e), type=ProxyErrorTypes.auth_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=status.HTTP_400_BAD_REQUEST, ) @@ -3099,7 +3100,7 @@ async def update_model( raise ProxyException( message=getattr(e, "detail", f"Authentication Error({e})"), type=ProxyErrorTypes.auth_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=getattr(e, "status_code", status.HTTP_400_BAD_REQUEST), ) elif isinstance(e, ProxyException): @@ -3107,7 +3108,7 @@ async def update_model( raise ProxyException( message="Authentication Error, " + str(e), type=ProxyErrorTypes.auth_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=status.HTTP_400_BAD_REQUEST, ) diff --git a/litellm/proxy/management_endpoints/organization_endpoints.py b/litellm/proxy/management_endpoints/organization_endpoints.py index e1d19d8efae..26e12c89b77 100644 --- a/litellm/proxy/management_endpoints/organization_endpoints.py +++ b/litellm/proxy/management_endpoints/organization_endpoints.py @@ -41,6 +41,7 @@ from litellm.proxy.auth.auth_checks import ( ) from litellm.proxy.auth.user_api_key_auth import user_api_key_auth from litellm.proxy.common_utils.auth_cache_invalidation_pubsub import evict_and_broadcast +from litellm.proxy.common_utils.openai_error_payload import openai_error_param from litellm.proxy.common_utils.timezone_utils import get_budget_reset_time from litellm.proxy.common_utils.user_api_key_cache import UserApiKeyCache from litellm.proxy.management_endpoints.budget_management_endpoints import ( @@ -1400,7 +1401,7 @@ async def organization_member_add( raise ProxyException( message=getattr(e, "detail", f"Authentication Error({e})"), type=ProxyErrorTypes.auth_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=getattr(e, "status_code", status.HTTP_500_INTERNAL_SERVER_ERROR), ) elif isinstance(e, ProxyException): @@ -1408,7 +1409,7 @@ async def organization_member_add( raise ProxyException( message="Authentication Error, " + str(e), type=ProxyErrorTypes.auth_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=status.HTTP_500_INTERNAL_SERVER_ERROR, ) diff --git a/litellm/proxy/management_endpoints/team_callback_endpoints.py b/litellm/proxy/management_endpoints/team_callback_endpoints.py index d1e1c4bd25e..5b778392664 100644 --- a/litellm/proxy/management_endpoints/team_callback_endpoints.py +++ b/litellm/proxy/management_endpoints/team_callback_endpoints.py @@ -46,6 +46,7 @@ from litellm.proxy.common_utils.callback_utils import ( # noqa: F401 # legacy encrypt_callback_vars, is_sensitive_callback_key, ) +from litellm.proxy.common_utils.openai_error_payload import openai_error_param from litellm.proxy.litellm_pre_call_utils import ( # noqa: F401 # legacy module exports _get_validated_callback_metadata, # pyright: ignore[reportPrivateUsage,reportUnusedImport] # backwards-compatible package export convert_key_logging_metadata_to_callback, @@ -463,7 +464,7 @@ async def add_team_callbacks( raise ProxyException( message="Internal Server Error, " + str(e), type=ProxyErrorTypes.internal_server_error.value, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=status.HTTP_500_INTERNAL_SERVER_ERROR, ) @@ -606,7 +607,7 @@ async def delete_team_callback( raise ProxyException( message="Internal Server Error, " + str(e), type=ProxyErrorTypes.internal_server_error.value, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=status.HTTP_500_INTERNAL_SERVER_ERROR, ) else: @@ -747,7 +748,7 @@ async def disable_team_logging( raise ProxyException( message="Internal Server Error, " + str(e), type=ProxyErrorTypes.internal_server_error.value, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=status.HTTP_500_INTERNAL_SERVER_ERROR, ) @@ -838,7 +839,7 @@ async def get_team_callbacks( raise ProxyException( message=getattr(e, "detail", f"Internal Server Error({e})"), type=ProxyErrorTypes.internal_server_error.value, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=getattr(e, "status_code", status.HTTP_500_INTERNAL_SERVER_ERROR), ) elif isinstance(e, ProxyException): @@ -846,6 +847,6 @@ async def get_team_callbacks( raise ProxyException( message="Internal Server Error, " + str(e), type=ProxyErrorTypes.internal_server_error.value, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=status.HTTP_500_INTERNAL_SERVER_ERROR, ) diff --git a/litellm/proxy/management_endpoints/team_endpoints.py b/litellm/proxy/management_endpoints/team_endpoints.py index 1fa18e9b2b8..513c3f720a0 100644 --- a/litellm/proxy/management_endpoints/team_endpoints.py +++ b/litellm/proxy/management_endpoints/team_endpoints.py @@ -131,6 +131,7 @@ from litellm.proxy.common_utils.auth_cache_invalidation_pubsub import ( ) from litellm.proxy.common_utils.callback_utils import encrypt_callback_vars from litellm.proxy.common_utils.json_merge_patch import apply_json_merge_patch +from litellm.proxy.common_utils.openai_error_payload import openai_error_param from litellm.proxy.common_utils.user_api_key_cache import ( UserApiKeyCache, team_membership_auth_cache_key, @@ -5454,7 +5455,7 @@ async def team_info( raise ProxyException( message=getattr(e, "detail", f"Authentication Error({e})"), type=ProxyErrorTypes.auth_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=getattr(e, "status_code", status.HTTP_400_BAD_REQUEST), ) elif isinstance(e, ProxyException): @@ -5462,7 +5463,7 @@ async def team_info( raise ProxyException( message="Authentication Error, " + str(e), type=ProxyErrorTypes.auth_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=status.HTTP_400_BAD_REQUEST, ) diff --git a/litellm/proxy/proxy_server.py b/litellm/proxy/proxy_server.py index ca0f8892b9d..bc2a94ce439 100644 --- a/litellm/proxy/proxy_server.py +++ b/litellm/proxy/proxy_server.py @@ -522,8 +522,11 @@ from litellm.proxy.common_utils.openai_endpoint_utils import ( ) from litellm.proxy.common_utils.openai_error_payload import ( LITELLM_CALL_ID_HEADER, + error_status_code, headers_with_litellm_call_id, litellm_call_id_headers, + openai_error_param, + openai_error_type, with_litellm_call_id, ) from litellm.proxy.common_utils.periodic_reload_schedule import ( @@ -9829,9 +9832,9 @@ async def async_assistants_data_generator(response, user_api_key_dict: UserAPIKe proxy_exception: Final = ProxyException( message=getattr(e, "message", error_msg), - type=getattr(e, "type", "None"), - param=getattr(e, "param", "None"), - code=getattr(e, "status_code", 500), + type=openai_error_type(e, error_status_code(e, 500)), + param=openai_error_param(e), + code=error_status_code(e, 500), ) error_returned: Final = json.dumps({"error": proxy_exception.to_dict()}) yield f"data: {error_returned}\n\n" @@ -10666,9 +10669,9 @@ async def async_data_generator( if isinstance(client_error, HTTPException) else ProxyException( message=getattr(client_error, "message", str(client_error)), - type=getattr(client_error, "type", "None"), - param=getattr(client_error, "param", "None"), - code=getattr(client_error, "status_code", 500), + type=openai_error_type(client_error, error_status_code(client_error, 500)), + param=openai_error_param(client_error), + code=error_status_code(client_error, 500), ).to_dict() }, default=str, @@ -13021,11 +13024,11 @@ async def completion( error_msg: Final = f"{e}" raise ProxyException( message=getattr(e, "message", error_msg), - type=getattr(e, "type", "None"), - param=getattr(e, "param", "None"), + type=openai_error_type(e, error_status_code(e, 500)), + param=openai_error_param(e), headers=litellm_call_id_headers(litellm_call_id), openai_code=getattr(e, "code", None), - code=getattr(e, "status_code", 500), + code=error_status_code(e, 500), ) @@ -13248,19 +13251,19 @@ async def moderations( if isinstance(e, HTTPException): raise ProxyException( message=getattr(e, "message", str(e)), - type=getattr(e, "type", "None"), - param=getattr(e, "param", "None"), + type=openai_error_type(e, error_status_code(e, status.HTTP_400_BAD_REQUEST)), + param=openai_error_param(e), headers=litellm_call_id_headers(litellm_call_id), - code=getattr(e, "status_code", status.HTTP_400_BAD_REQUEST), + code=error_status_code(e, status.HTTP_400_BAD_REQUEST), ) else: error_msg: Final = f"{e}" raise ProxyException( message=getattr(e, "message", error_msg), - type=getattr(e, "type", "None"), - param=getattr(e, "param", "None"), + type=openai_error_type(e, error_status_code(e, 500)), + param=openai_error_param(e), headers=litellm_call_id_headers(litellm_call_id), - code=getattr(e, "status_code", 500), + code=error_status_code(e, 500), ) @@ -13401,11 +13404,11 @@ async def audio_speech( ) raise ProxyException( message=getattr(e, "message", f"{e}"), - type=getattr(e, "type", "None"), - param=getattr(e, "param", "None"), + type=openai_error_type(e, error_status_code(e, 500)), + param=openai_error_param(e), headers=litellm_call_id_headers(litellm_call_id), openai_code=getattr(e, "code", None), - code=getattr(e, "status_code", 500), + code=error_status_code(e, 500), ) @@ -13547,20 +13550,20 @@ async def audio_transcriptions( if isinstance(e, HTTPException): raise ProxyException( message=getattr(e, "message", str(e.detail)), - type=getattr(e, "type", "None"), - param=getattr(e, "param", "None"), + type=openai_error_type(e, error_status_code(e, status.HTTP_400_BAD_REQUEST)), + param=openai_error_param(e), headers=litellm_call_id_headers(litellm_call_id), - code=getattr(e, "status_code", status.HTTP_400_BAD_REQUEST), + code=error_status_code(e, status.HTTP_400_BAD_REQUEST), ) else: error_msg: Final = f"{e}" raise ProxyException( message=getattr(e, "message", error_msg), - type=getattr(e, "type", "None"), - param=getattr(e, "param", "None"), + type=openai_error_type(e, error_status_code(e, 500)), + param=openai_error_param(e), headers=litellm_call_id_headers(litellm_call_id), openai_code=getattr(e, "code", None), - code=getattr(e, "status_code", 500), + code=error_status_code(e, 500), ) @@ -17071,7 +17074,7 @@ async def model_streaming_metrics( raise ProxyException( message=CommonProxyErrors.db_not_connected_error.value, type="internal_error", - param="None", + param=None, code=status.HTTP_500_INTERNAL_SERVER_ERROR, ) @@ -17205,7 +17208,7 @@ async def model_metrics( raise ProxyException( message="Prisma Client is not initialized", type="internal_error", - param="None", + param=None, code=status.HTTP_500_INTERNAL_SERVER_ERROR, ) startTime = startTime or datetime.now() - timedelta(days=DAYS_IN_A_MONTH) @@ -17320,7 +17323,7 @@ async def model_metrics_slow_responses( raise ProxyException( message="Prisma Client is not initialized", type="internal_error", - param="None", + param=None, code=status.HTTP_500_INTERNAL_SERVER_ERROR, ) if api_key is None or api_key == "undefined": @@ -17409,7 +17412,7 @@ async def model_metrics_exceptions( raise ProxyException( message="Prisma Client is not initialized", type="internal_error", - param="None", + param=None, code=status.HTTP_500_INTERNAL_SERVER_ERROR, ) @@ -18439,7 +18442,7 @@ async def async_queue_request( raise ProxyException( message=getattr(e, "detail", f"Authentication Error({e})"), type=ProxyErrorTypes.auth_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=getattr(e, "status_code", status.HTTP_400_BAD_REQUEST), ) elif isinstance(e, ProxyException): @@ -18451,7 +18454,7 @@ async def async_queue_request( raise ProxyException( message="Authentication Error, " + str(e), type=ProxyErrorTypes.auth_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=status.HTTP_400_BAD_REQUEST, ) @@ -18626,7 +18629,7 @@ async def login_v2(request: Request): raise ProxyException( message=getattr(e, "detail", str(e)), type=ProxyErrorTypes.auth_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=getattr(e, "status_code", status.HTTP_500_INTERNAL_SERVER_ERROR), ) else: @@ -18634,7 +18637,7 @@ async def login_v2(request: Request): raise ProxyException( message=error_msg, type=ProxyErrorTypes.auth_error, - param="None", + param=None, code=status.HTTP_500_INTERNAL_SERVER_ERROR, ) @@ -18702,7 +18705,7 @@ async def login_v3(request: Request): raise ProxyException( message=getattr(e, "detail", str(e)), type=ProxyErrorTypes.auth_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=getattr(e, "status_code", status.HTTP_500_INTERNAL_SERVER_ERROR), ) else: @@ -18710,7 +18713,7 @@ async def login_v3(request: Request): raise ProxyException( message=error_msg, type=ProxyErrorTypes.auth_error, - param="None", + param=None, code=status.HTTP_500_INTERNAL_SERVER_ERROR, ) @@ -18775,7 +18778,7 @@ async def login_v3_exchange(request: Request): raise ProxyException( message=str(e), type=ProxyErrorTypes.auth_error, - param="None", + param=None, code=status.HTTP_500_INTERNAL_SERVER_ERROR, ) @@ -19736,7 +19739,7 @@ async def update_config( raise ProxyException( message=getattr(e, "detail", f"Authentication Error({e})"), type=ProxyErrorTypes.auth_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=getattr(e, "status_code", status.HTTP_400_BAD_REQUEST), ) elif isinstance(e, ProxyException): @@ -19744,7 +19747,7 @@ async def update_config( raise ProxyException( message="Authentication Error, " + str(e), type=ProxyErrorTypes.auth_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=status.HTTP_400_BAD_REQUEST, ) @@ -20931,7 +20934,7 @@ async def get_config( raise ProxyException( message=getattr(e, "detail", f"Authentication Error({e})"), type=ProxyErrorTypes.auth_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=getattr(e, "status_code", status.HTTP_400_BAD_REQUEST), ) elif isinstance(e, ProxyException): @@ -20939,7 +20942,7 @@ async def get_config( raise ProxyException( message="Authentication Error, " + str(e), type=ProxyErrorTypes.auth_error, - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=status.HTTP_400_BAD_REQUEST, ) diff --git a/litellm/proxy/spend_tracking/spend_management_endpoints.py b/litellm/proxy/spend_tracking/spend_management_endpoints.py index 3c65fc0bf06..db91418a6e0 100644 --- a/litellm/proxy/spend_tracking/spend_management_endpoints.py +++ b/litellm/proxy/spend_tracking/spend_management_endpoints.py @@ -50,6 +50,7 @@ from litellm.proxy.auth.authorization_dependencies import ( LogTeamLookupDependency, ) from litellm.proxy.auth.user_api_key_auth import user_api_key_auth +from litellm.proxy.common_utils.openai_error_payload import error_status_code, openai_error_param, openai_error_type from litellm.proxy.litellm_pre_call_utils import LiteLLMProxyRequestSetup from litellm.proxy.spend_tracking.spend_capture_rate import ( ProviderBillingCredentialMissing, @@ -636,7 +637,7 @@ async def view_spend_tags( raise ProxyException( message=getattr(e, "detail", f"/spend/tags Error({e})"), type="internal_error", - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=getattr(e, "status_code", status.HTTP_500_INTERNAL_SERVER_ERROR), ) elif isinstance(e, ProxyException): @@ -644,7 +645,7 @@ async def view_spend_tags( raise ProxyException( message="/spend/tags Error" + str(e), type="internal_error", - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=status.HTTP_500_INTERNAL_SERVER_ERROR, ) @@ -2142,7 +2143,7 @@ async def global_get_all_tag_names(): raise ProxyException( message=getattr(e, "detail", f"/spend/all_tag_names Error({e})"), type="internal_error", - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=getattr(e, "status_code", status.HTTP_500_INTERNAL_SERVER_ERROR), ) elif isinstance(e, ProxyException): @@ -2150,7 +2151,7 @@ async def global_get_all_tag_names(): raise ProxyException( message="/spend/all_tag_names Error" + str(e), type="internal_error", - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=status.HTTP_500_INTERNAL_SERVER_ERROR, ) @@ -2224,7 +2225,7 @@ async def global_view_spend_tags( raise ProxyException( message=getattr(e, "detail", f"/spend/tags Error({error_str})"), type="internal_error", - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=getattr(e, "status_code", status.HTTP_500_INTERNAL_SERVER_ERROR), ) elif isinstance(e, ProxyException): @@ -2232,7 +2233,7 @@ async def global_view_spend_tags( raise ProxyException( message="/spend/tags Error" + error_str, type="internal_error", - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=status.HTTP_500_INTERNAL_SERVER_ERROR, ) @@ -2445,9 +2446,9 @@ async def calculate_spend(request: SpendCalculateRequest): if isinstance(e, HTTPException): raise ProxyException( message=getattr(e, "detail", str(e)), - type=getattr(e, "type", "None"), - param=getattr(e, "param", "None"), - code=getattr(e, "status_code", status.HTTP_400_BAD_REQUEST), + type=openai_error_type(e, error_status_code(e, status.HTTP_400_BAD_REQUEST)), + param=openai_error_param(e), + code=error_status_code(e, status.HTTP_400_BAD_REQUEST), ) if isinstance(e, litellm.exceptions.ModelNotMappedError): raise ProxyException( @@ -2459,9 +2460,9 @@ async def calculate_spend(request: SpendCalculateRequest): error_msg: Final = f"{e}" raise ProxyException( message=getattr(e, "message", error_msg), - type=getattr(e, "type", "None"), - param=getattr(e, "param", "None"), - code=getattr(e, "status_code", 500), + type=openai_error_type(e, error_status_code(e, 500)), + param=openai_error_param(e), + code=error_status_code(e, 500), ) @@ -2631,7 +2632,7 @@ async def ui_view_spend_logs( raise ProxyException( message="Prisma Client is not initialized", type="internal_error", - param="None", + param=None, code=status.HTTP_401_UNAUTHORIZED, ) @@ -2698,7 +2699,7 @@ async def ui_view_spend_logs( raise ProxyException( message="Start date and end date are required", type="bad_request", - param="None", + param=None, code=status.HTTP_400_BAD_REQUEST, ) formats: Final = ["%Y-%m-%d %H:%M:%S", "%Y-%m-%d"] if is_v2 else ["%Y-%m-%d %H:%M:%S"] @@ -3679,7 +3680,7 @@ async def view_spend_logs( raise ProxyException( message=getattr(e, "detail", f"/spend/logs Error({e})"), type="internal_error", - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=getattr(e, "status_code", status.HTTP_500_INTERNAL_SERVER_ERROR), ) elif isinstance(e, ProxyException): @@ -3687,7 +3688,7 @@ async def view_spend_logs( raise ProxyException( message="/spend/logs Error" + str(e), type="internal_error", - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=status.HTTP_500_INTERNAL_SERVER_ERROR, ) @@ -3714,7 +3715,7 @@ async def global_spend_reset(): raise ProxyException( message="Prisma Client is not initialized", type="internal_error", - param="None", + param=None, code=status.HTTP_401_UNAUTHORIZED, ) @@ -3745,7 +3746,7 @@ async def global_spend_refresh(): raise ProxyException( message="Prisma Client is not initialized", type="internal_error", - param="None", + param=None, code=status.HTTP_401_UNAUTHORIZED, ) @@ -3816,7 +3817,7 @@ async def global_spend_for_internal_user( raise ProxyException( message="Prisma Client is not initialized", type="internal_error", - param="None", + param=None, code=status.HTTP_500_INTERNAL_SERVER_ERROR, ) try: @@ -3877,7 +3878,7 @@ async def global_spend_logs( raise ProxyException( message="Prisma Client is not initialized", type="internal_error", - param="None", + param=None, code=status.HTTP_500_INTERNAL_SERVER_ERROR, ) @@ -3921,7 +3922,7 @@ async def global_spend_logs( raise ProxyException( message=getattr(e, "detail", f"/global/spend/logs Error({error_str})"), type="internal_error", - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=getattr(e, "status_code", status.HTTP_500_INTERNAL_SERVER_ERROR), ) elif isinstance(e, ProxyException): @@ -3929,7 +3930,7 @@ async def global_spend_logs( raise ProxyException( message="/global/spend/logs Error" + error_str, type="internal_error", - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=status.HTTP_500_INTERNAL_SERVER_ERROR, ) @@ -3969,7 +3970,7 @@ async def global_spend(): raise ProxyException( message=getattr(e, "detail", f"/global/spend Error({error_str})"), type="internal_error", - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=getattr(e, "status_code", status.HTTP_500_INTERNAL_SERVER_ERROR), ) elif isinstance(e, ProxyException): @@ -3977,7 +3978,7 @@ async def global_spend(): raise ProxyException( message="/global/spend Error" + error_str, type="internal_error", - param=getattr(e, "param", "None"), + param=openai_error_param(e), code=status.HTTP_500_INTERNAL_SERVER_ERROR, ) diff --git a/tests/unit/proxy/auth/test_auth_exception_handler.py b/tests/unit/proxy/auth/test_auth_exception_handler.py index e21491fd996..007fdd8192d 100644 --- a/tests/unit/proxy/auth/test_auth_exception_handler.py +++ b/tests/unit/proxy/auth/test_auth_exception_handler.py @@ -1285,3 +1285,17 @@ def test_as_proxy_exception_keeps_jwt_scope_denial_message_shape(): assert converted.to_dict() == plain.to_dict() assert converted.internal_message == denial.internal_message + + +@pytest.mark.parametrize( + "auth_failure", + [ + pytest.param(HTTPException(status_code=401, detail="Authentication Error, invalid key"), id="http_exception"), + pytest.param(Exception("Invalid proxy server token passed"), id="bare_exception"), + ], +) +def test_auth_failure_error_body_carries_a_json_null_param(auth_failure: Exception) -> None: + body = json.loads(json.dumps({"error": _as_proxy_exception(auth_failure).to_dict()})) + + assert body["error"]["type"] == ProxyErrorTypes.auth_error.value + assert body["error"]["param"] is None diff --git a/tests/unit/proxy/auth/test_handle_jwt.py b/tests/unit/proxy/auth/test_handle_jwt.py index a0ea6ce0403..51cf15996cd 100644 --- a/tests/unit/proxy/auth/test_handle_jwt.py +++ b/tests/unit/proxy/auth/test_handle_jwt.py @@ -4368,6 +4368,7 @@ async def test_removed_signing_key_stops_being_trusted_once_the_stale_window_exp assert exc_info.value.code == "503" assert exc_info.value.type == ProxyErrorTypes.auth_provider_unavailable + assert exc_info.value.param is None @pytest.mark.asyncio diff --git a/tests/unit/proxy/health_endpoints/test_health_endpoints.py b/tests/unit/proxy/health_endpoints/test_health_endpoints.py index 67b280eb379..3bb65fdafd2 100644 --- a/tests/unit/proxy/health_endpoints/test_health_endpoints.py +++ b/tests/unit/proxy/health_endpoints/test_health_endpoints.py @@ -1710,6 +1710,18 @@ async def test_health_services_endpoint_rejects_unknown_service(): await health_services_endpoint(service="totally_unknown_service_xyz") +@pytest.mark.asyncio +async def test_health_services_endpoint_rejection_is_openai_shaped(): + """A caller mistake here must serialize as an OpenAI error object: a real + `type` string and a JSON null `param`, never the literal string "None".""" + with pytest.raises(ProxyException) as exc_info: + await health_services_endpoint(service="totally_unknown_service_xyz") + + body = json.loads(json.dumps({"error": exc_info.value.to_dict()})) + assert body["error"]["type"] == "auth_error" + assert body["error"]["param"] is None + + @pytest.mark.asyncio @pytest.mark.parametrize( "role", diff --git a/tests/unit/proxy/management_endpoints/test_internal_user_endpoints.py b/tests/unit/proxy/management_endpoints/test_internal_user_endpoints.py index cb8a73d581b..7dcfaba35d2 100644 --- a/tests/unit/proxy/management_endpoints/test_internal_user_endpoints.py +++ b/tests/unit/proxy/management_endpoints/test_internal_user_endpoints.py @@ -4970,7 +4970,7 @@ _DB_OUTAGE_503_BODY: Final = { "error": { "message": "Service Unavailable, the authentication database is temporarily unreachable. Please retry shortly.", "type": "no_db_connection", - "param": "None", + "param": None, "code": "503", } } diff --git a/tests/unit/proxy/management_endpoints/test_key_management_endpoints.py b/tests/unit/proxy/management_endpoints/test_key_management_endpoints.py index d6ee9b1d29e..fcf9f5113cf 100644 --- a/tests/unit/proxy/management_endpoints/test_key_management_endpoints.py +++ b/tests/unit/proxy/management_endpoints/test_key_management_endpoints.py @@ -23306,4 +23306,18 @@ def test_key_info_database_not_connected_sanitized_500( proxy_server.app.dependency_overrides.pop(user_api_key_auth, None) +def test_key_health_failure_body_carries_a_json_null_param(): + import litellm.proxy.proxy_server as ps + app.dependency_overrides[ps.user_api_key_auth] = lambda: UserAPIKeyAuth( + api_key="hashed-key", metadata={"logging": ["not-a-callback-object"]} + ) + try: + response = client.post("/key/health", headers={"Authorization": "Bearer sk-test"}) + finally: + app.dependency_overrides.pop(ps.user_api_key_auth, None) + + assert response.status_code == 500 + error = response.json()["error"] + assert error["type"] == "internal_server_error" + assert error["param"] is None diff --git a/tests/unit/proxy/management_endpoints/test_team_endpoints.py b/tests/unit/proxy/management_endpoints/test_team_endpoints.py index f727448f3ac..9adfb64c402 100644 --- a/tests/unit/proxy/management_endpoints/test_team_endpoints.py +++ b/tests/unit/proxy/management_endpoints/test_team_endpoints.py @@ -18827,7 +18827,7 @@ _DB_OUTAGE_503_BODY: Final = { "error": { "message": "Service Unavailable, the authentication database is temporarily unreachable. Please retry shortly.", "type": "no_db_connection", - "param": "None", + "param": None, "code": "503", } } diff --git a/tests/unit/proxy/proxy_server/test_exception_handlers.py b/tests/unit/proxy/proxy_server/test_exception_handlers.py index f6efdecc884..40a07890f15 100644 --- a/tests/unit/proxy/proxy_server/test_exception_handlers.py +++ b/tests/unit/proxy/proxy_server/test_exception_handlers.py @@ -438,7 +438,7 @@ _DB_OUTAGE_503_BODY: Final = { "error": { "message": "Service Unavailable, the authentication database is temporarily unreachable. Please retry shortly.", "type": "no_db_connection", - "param": "None", + "param": None, "code": "503", } } diff --git a/tests/unit/proxy/spend_tracking/test_spend_management_endpoints.py b/tests/unit/proxy/spend_tracking/test_spend_management_endpoints.py index 8d53817f51e..6598b94e60f 100644 --- a/tests/unit/proxy/spend_tracking/test_spend_management_endpoints.py +++ b/tests/unit/proxy/spend_tracking/test_spend_management_endpoints.py @@ -8268,3 +8268,19 @@ async def test_global_spend_refresh_runs_the_refresh_on_the_writer_when_a_reader assert [statement for statement in journal.statements if statement[1].startswith("REFRESH")] == [ ("writer", 'REFRESH MATERIALIZED VIEW "MonthlyGlobalSpend";') ] + + +def test_spend_calculate_rejection_body_is_openai_shaped(client): + app.dependency_overrides[ps.user_api_key_auth] = lambda: UserAPIKeyAuth( + user_role=LitellmUserRoles.PROXY_ADMIN, user_id="admin_user", api_key="hashed-k" + ) + try: + response = client.post("/spend/calculate", json={}, headers={"Authorization": "Bearer sk-test"}) + finally: + app.dependency_overrides.pop(ps.user_api_key_auth, None) + + assert response.status_code == 400 + error = response.json()["error"] + assert error["type"] == "invalid_request_error" + assert error["param"] is None + assert error["code"] == "400" diff --git a/tests/unit/proxy/test_proxy_server_endpoints_and_startup.py b/tests/unit/proxy/test_proxy_server_endpoints_and_startup.py index 745488bf93f..6beb837de8a 100644 --- a/tests/unit/proxy/test_proxy_server_endpoints_and_startup.py +++ b/tests/unit/proxy/test_proxy_server_endpoints_and_startup.py @@ -7250,6 +7250,39 @@ async def test_async_data_generator_closes_the_guardrail_chain_before_returning_ assert cleanup_ran == [True] +@pytest.mark.asyncio +async def test_async_data_generator_midstream_failure_frame_is_openai_shaped_with_a_json_null_param(): + from litellm.proxy._types import UserAPIKeyAuth + from litellm.proxy.proxy_server import async_data_generator + + async def chain_failing_after_first_chunk(**_kwargs): + yield {"choices": [{"delta": {"content": "Hello"}}]} + raise ValueError("audit hook could not reach its sink") + + proxy_logging = _guarded_chain_logging(chain_failing_after_first_chunk) + proxy_logging.post_call_failure_hook = AsyncMock(return_value=None) + + with patch("litellm.proxy.proxy_server.proxy_logging_obj", proxy_logging): + frames = [ + frame + async for frame in async_data_generator( + MagicMock(), MagicMock(spec=UserAPIKeyAuth), {"model": "gpt-4o-mini"} + ) + ] + + error_bodies = [json.loads(frame.removeprefix("data: ")) for frame in frames if '"error"' in frame] + assert error_bodies == [ + { + "error": { + "message": "audit hook could not reach its sink", + "type": "internal_server_error", + "param": None, + "code": "500", + } + } + ] + + @pytest.mark.asyncio async def test_async_data_generator_closes_the_guardrail_chain_while_a_keepalive_read_is_pending(): from litellm.proxy._types import UserAPIKeyAuth @@ -14217,6 +14250,51 @@ async def test_moderations_reraises_proxy_exception_unwrapped(): mock_logging.post_call_failure_hook.assert_awaited_once() +@pytest.mark.asyncio +@pytest.mark.parametrize( + ("failure", "expected_type", "expected_code"), + [ + pytest.param(HTTPException(status_code=400, detail="bad input"), "invalid_request_error", "400", id="http_400"), + pytest.param(HTTPException(status_code=429, detail="slow down"), "rate_limit_error", "429", id="http_429"), + pytest.param(Exception("provider down"), "internal_server_error", "500", id="bare_exception"), + ], +) +async def test_moderations_failure_body_is_openai_shaped(failure, expected_type, expected_code): + from litellm.proxy._types import ProxyException + + async def passthrough_add_litellm_data(data, **kwargs): + return data + + request = MagicMock() + request.body = AsyncMock(return_value=b'{"input": "hi"}') + fake_logging = MagicMock() + fake_logging.pre_call_hook = AsyncMock(side_effect=lambda user_api_key_dict, data, call_type: data) + fake_logging.post_call_failure_hook = AsyncMock() + + with ( + patch.object( + proxy_server_module, "add_litellm_data_to_request", new=passthrough_add_litellm_data + ), # test-quality-ok: the route reads this module global, no injection point + patch.object( + proxy_server_module, "route_request", new=AsyncMock(side_effect=failure) + ), # test-quality-ok: fakes the provider failure so the real route's error shaping is observable + patch.object( + proxy_server_module, "proxy_logging_obj", new=fake_logging + ), # test-quality-ok: module global, no injection point + pytest.raises(ProxyException) as raised, + ): + await proxy_server_module.moderations( + request=request, + fastapi_response=MagicMock(), + user_api_key_dict=UserAPIKeyAuth(api_key="sk-test"), + ) + + body = json.loads(json.dumps({"error": raised.value.to_dict()})) + assert body["error"]["type"] == expected_type + assert body["error"]["param"] is None + assert body["error"]["code"] == expected_code + + @pytest.mark.asyncio async def test_moderations_response_carries_litellm_call_id_header(): from fastapi import Response