mirror of
https://github.com/BerriAI/litellm.git
synced 2026-09-17 23:51:30 +00:00
c6f77aa964
OpenAI-compatible endpoints parse `x-litellm-spend-logs-metadata` in `LiteLLMProxyRequestSetup._get_spend_logs_metadata_from_request_headers` and place it at `data["metadata"]["spend_logs_metadata"]`. Pass-through endpoints (/anthropic/v1/messages, /openai/v1/*, /v2/chat, /vertex_ai/*, /gemini/*, Cursor) bypass that pipeline, so `StandardLoggingPayload.metadata.spend_logs_metadata` was always None for `call_type=pass_through_endpoint`, hiding per-call metadata (e.g. task / issue) from `/spend/logs`. This change extracts the header at both `PassthroughStandardLoggingPayload` construction sites (HTTP and WebSocket), adds an optional `spend_logs_metadata` field to the TypedDict, and propagates it into `kwargs["litellm_params"]["metadata"]` at every provider-handler payload builder (base, Anthropic, OpenAI, Cohere, Gemini, Vertex) — both the non-streaming and streaming (collected-chunks) paths. Two small shared helpers on `BasePassthroughLoggingHandler` keep the propagation consistent across handlers: - `_apply_spend_logs_metadata(kwargs, payload)` — writes to the same key path OpenAI-compatible flows already use, so `get_standard_logging_object_payload` surfaces the value identically. - `_seed_streaming_kwargs_from_logging_obj(logging_obj)` — seeds the fresh kwargs used by streaming logging paths with the payload that `success_handler.pass_through_async_success_handler` stashes on `model_call_details` before any provider handler runs. Tests in `tests/test_litellm/proxy/pass_through_endpoints/test_spend_logs_metadata_propagation.py` cover: header extraction (valid / missing / malformed JSON), helper writes to the matching key path, preservation of existing metadata keys, no-op when payload or metadata is absent, streaming-kwargs seed behavior. |
||
|---|---|---|
| .. | ||
| agent_tests | ||
| audio_tests | ||
| basic_proxy_startup_tests | ||
| batches_tests | ||
| benchmarks | ||
| code_coverage_tests | ||
| documentation_tests | ||
| enterprise | ||
| guardrails_tests | ||
| image_gen_tests | ||
| litellm | ||
| litellm-proxy-extras | ||
| litellm_core_utils | ||
| litellm_utils_tests | ||
| llm_responses_api_testing | ||
| llm_translation | ||
| load_tests | ||
| local_testing | ||
| logging_callback_tests | ||
| mcp_tests | ||
| multi_instance_e2e_tests | ||
| ocr_tests | ||
| old_proxy_tests/tests | ||
| openai_endpoints_tests | ||
| otel_tests | ||
| pass_through_tests | ||
| pass_through_unit_tests | ||
| proxy_admin_ui_tests | ||
| proxy_e2e_anthropic_messages_tests | ||
| proxy_security_tests | ||
| proxy_unit_tests | ||
| router_unit_tests | ||
| scim_tests | ||
| search_tests | ||
| spend_tracking_tests | ||
| store_model_in_db_tests | ||
| test_litellm | ||
| unified_google_tests | ||
| vector_store_tests | ||
| windows_tests | ||
| __init__.py | ||
| gettysburg.wav | ||
| large_text.py | ||
| openai_batch_completions.jsonl | ||
| README.MD | ||
| test_budget_management.py | ||
| test_callbacks_on_proxy.py | ||
| test_config.py | ||
| test_debug_warning.py | ||
| test_default_encoding_non_root.py | ||
| test_end_users.py | ||
| test_entrypoint.py | ||
| test_fallbacks.py | ||
| test_gpt5_azure_temperature_support.py | ||
| test_health.py | ||
| test_keys.py | ||
| test_litellm_proxy_responses_config.py | ||
| test_logging.conf | ||
| test_models.py | ||
| test_new_vector_store_endpoints.py | ||
| test_openai_endpoints.py | ||
| test_organizations.py | ||
| test_otel_thread_leak.py | ||
| test_passthrough_endpoints.py | ||
| test_presidio_latency.py | ||
| test_proxy_server_non_root.py | ||
| test_ratelimit.py | ||
| test_resource_cleanup.py | ||
| test_service_logger_otel.py | ||
| test_spend_logs.py | ||
| test_team.py | ||
| test_team_logging.py | ||
| test_team_members.py | ||
| test_users.py | ||
In total litellm runs 1000+ tests
[02/20/2025] Update:
To make it easier to contribute and map what behavior is tested,
we've started mapping the litellm directory in tests/test_litellm
This folder can only run mock tests.