mirror of
https://github.com/BerriAI/litellm.git
synced 2026-09-06 08:16:43 +00:00
5866c862be
Fixes #24833 — _registered_pass_through_routes grows unbounded causing CPU to reach 100% when pass-through endpoints are stored in the DB. Two root-cause bugs: 1. **UUID churn** — DB-sourced endpoints have no uid=1000(ubuntu) gid=1000(ubuntu) groups=1000(ubuntu),27(sudo) field, so generated a fresh on every 30-second reload cycle. Each new UUID produced a different route key, so old entries never matched during cleanup and accumulated forever. With N endpoints and a 30-second cycle, the dict grew by N entries every 30 s and the O(n²) scan in consumed all available CPU. Fix: derive a stable, deterministic ID from the endpoint's path + methods using SHA-256 (first 16 hex chars, prefixed with 'auto-'). An unchanged endpoint now always maps to the same registry key across reloads. 2. **Wrong argument to remove_endpoint_routes** — the cleanup loop in passed the full route key (e.g. 'uuid:exact:/path:GET,POST') to , which expects only the endpoint_id prefix. The inner linear scan compared value['endpoint_id'] == full_route_key, which never matched any stored entry. Cleanup was a complete no-op regardless of whether the ID changed. Fix: extract the endpoint_id by splitting the route key on ':' with maxsplit=1 before calling remove_endpoint_routes(). Also added at module level (previously would have been a repeated inline import). Tests: 5 new unit tests in tests/pass_through_unit_tests/test_passthrough_registry_leak.py covering: - stable ID for same path across reloads - different paths produce different IDs - remove_endpoint_routes works when called with extracted endpoint_id - registry size does not grow across 5 consecutive reload cycles - explicit IDs are preserved (no regression) |
||
|---|---|---|
| .. | ||
| 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_e2e_azure_batches_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.