diff --git a/litellm/router_utils/pre_call_checks/encrypted_content_affinity_check.py b/litellm/router_utils/pre_call_checks/encrypted_content_affinity_check.py index b623e31ce06..eba2f36b675 100644 --- a/litellm/router_utils/pre_call_checks/encrypted_content_affinity_check.py +++ b/litellm/router_utils/pre_call_checks/encrypted_content_affinity_check.py @@ -37,6 +37,7 @@ Safe to enable globally: """ import time +from collections.abc import Mapping from typing import TYPE_CHECKING, Final, Optional, Protocol, cast import httpx @@ -48,6 +49,7 @@ from litellm.exceptions import ( ServiceUnavailableError, ) from litellm.integrations.custom_logger import CustomLogger, Span +from litellm.litellm_core_utils.credential_accessor import CredentialAccessor from litellm.responses.utils import ResponsesAPIRequestUtils from litellm.router_utils.cooldown_cache import CooldownCacheValue from litellm.types.llms.openai import AllMessageValues @@ -161,11 +163,13 @@ class EncryptedContentAffinityCheck(CustomLogger): @staticmethod def _encryption_boundary_key( litellm_params: object, - ) -> tuple | None: + ) -> tuple[object, object] | None: """ ``(api_base, api_key)`` pair identifying an Azure resource. Two deployments sharing both are interchangeable for ``encrypted_content`` follow-ups; Azure rejects content produced by any other resource. + Missing values are resolved from ``litellm_credential_name`` without + modifying the deployment, and explicit deployment values take precedence. Accepts any object exposing dict-style ``.get(key, default)``: plain dicts (the common case in ``healthy_deployments``) as well as @@ -180,9 +184,29 @@ class EncryptedContentAffinityCheck(CustomLogger): return None api_base: Final = getter("api_base") api_key: Final = getter("api_key") - if not api_base or not api_key: + credential_name: Final = getter("litellm_credential_name") + credential_values: Final[Mapping[str, object] | None] = ( + CredentialAccessor.get_credential_values(credential_name) + if isinstance(credential_name, str) and credential_name and (api_base is None or api_key is None) + else None + ) + effective_api_base: Final = ( + api_base + if api_base is not None + else credential_values.get("api_base") + if credential_values is not None + else None + ) + effective_api_key: Final = ( + api_key + if api_key is not None + else credential_values.get("api_key") + if credential_values is not None + else None + ) + if not effective_api_base or not effective_api_key: return None - return (api_base, api_key) + return (effective_api_base, effective_api_key) def _find_deployments_on_same_encryption_boundary( self, diff --git a/tests/test_litellm/router_utils/pre_call_checks/test_encrypted_content_affinity_check.py b/tests/test_litellm/router_utils/pre_call_checks/test_encrypted_content_affinity_check.py index dac991a41c4..46e67983c33 100644 --- a/tests/test_litellm/router_utils/pre_call_checks/test_encrypted_content_affinity_check.py +++ b/tests/test_litellm/router_utils/pre_call_checks/test_encrypted_content_affinity_check.py @@ -21,8 +21,8 @@ from unittest.mock import AsyncMock, patch import pytest - import litellm +from litellm.models.credentials import CredentialItem from litellm.responses.utils import ResponsesAPIRequestUtils from litellm.types.llms.openai import ResponsesAPIResponse @@ -1148,6 +1148,140 @@ def test_boundary_key_accepts_pydantic_litellm_params_instance(): ) +def test_boundary_key_resolves_missing_values_from_named_credential(): + from litellm.router_utils.pre_call_checks.encrypted_content_affinity_check import ( + EncryptedContentAffinityCheck, + ) + + with ( + patch.object( # test-quality-ok: credential registry is the direct dependency under test + litellm, + "credential_list", + [ + CredentialItem( + credential_name="account-a", + credential_values={ + "api_base": "https://account-a.example.com", + "api_key": "credential-key-a", + }, + credential_info={}, + ) + ], + ) + ): + boundary = EncryptedContentAffinityCheck._encryption_boundary_key({"litellm_credential_name": "account-a"}) + + assert boundary == ("https://account-a.example.com", "credential-key-a") + + +def test_boundary_key_prefers_explicit_values_over_named_credential(): + from litellm.router_utils.pre_call_checks.encrypted_content_affinity_check import ( + EncryptedContentAffinityCheck, + ) + + with ( + patch.object( # test-quality-ok: credential registry is the direct dependency under test + litellm, + "credential_list", + [ + CredentialItem( + credential_name="account-a", + credential_values={ + "api_base": "https://credential.example.com", + "api_key": "credential-key-a", + }, + credential_info={}, + ) + ], + ) + ): + boundary = EncryptedContentAffinityCheck._encryption_boundary_key( + { + "api_base": "https://deployment.example.com", + "litellm_credential_name": "account-a", + } + ) + + assert boundary == ("https://deployment.example.com", "credential-key-a") + + +def test_boundary_fallback_matches_deployments_with_same_named_credential_values(): + from litellm.router_utils.pre_call_checks.encrypted_content_affinity_check import ( + EncryptedContentAffinityCheck, + ) + + with ( + patch.object( # test-quality-ok: credential registry is the direct dependency under test + litellm, + "credential_list", + [ + CredentialItem( + credential_name="account-a", + credential_values={ + "api_base": "https://account-a.example.com", + "api_key": "credential-key-a", + }, + credential_info={}, + ), + CredentialItem( + credential_name="account-a-peer", + credential_values={ + "api_base": "https://account-a.example.com", + "api_key": "credential-key-a", + }, + credential_info={}, + ), + CredentialItem( + credential_name="account-b", + credential_values={ + "api_base": "https://account-b.example.com", + "api_key": "credential-key-b", + }, + credential_info={}, + ), + ], + ) + ): + router = litellm.Router( + model_list=[ + { + "model_name": "gpt-5.3-codex", + "litellm_params": { + "model": "azure/gpt-5.3-codex", + "litellm_credential_name": "account-a", + }, + "model_info": {"id": "origin"}, + } + ], + num_retries=0, + ) + check = EncryptedContentAffinityCheck(router=router) + healthy_deployments = [ + { + "model_info": {"id": "peer-same-boundary"}, + "litellm_params": { + "model": "azure/gpt-5.4", + "litellm_credential_name": "account-a-peer", + }, + }, + { + "model_info": {"id": "peer-different-boundary"}, + "litellm_params": { + "model": "azure/gpt-5.4", + "litellm_credential_name": "account-b", + }, + }, + ] + + matches, originating = check._find_deployments_on_same_encryption_boundary( + healthy_deployments=healthy_deployments, + model_id="origin", + ) + + assert originating is not None + assert [deployment["model_info"]["id"] for deployment in matches] == ["peer-same-boundary"] + + def test_boundary_key_rejects_non_dict_like_inputs(): """ Inputs that don't expose ``.get()`` (None, lists, strings, ints) -> None.