* test: drop the cwd-relative sys.path.insert calls from the test suite
TQ003 stands at 1,077 across 1,058 files, and 1,015 of them are the same shape:
sys.path.insert(0, os.path.abspath("../..")) and its deeper siblings. The
argument resolves against the working directory rather than the file, so from
the repo root, where every job runs pytest, it inserts the directory two levels
above the checkout. It has never pointed at litellm. The package is installed
into the environment anyway, which is what actually makes the import work, and
what the rule's message has said all along.
Removing them leaves 1,634 imports of sys and os with no remaining reference,
and those go too, except where another test module imports the name back out of
the file. The rest of TQ003 is 62 call sites that resolve against __file__ or a
variable, which are a different question and are left alone.
Collection is identical either way: 45,871 tests and the same 51 pre-existing
collection errors before and after, and ruff reports no new undefined name.
* test: drop the duplicate imports the sys.path sweep exposed to F811
* test(pre-call-utils): restore the os import the new bedrock tests need
* fix(test): add missing mocks for test_streamable_http_mcp_handler_mock
The test was missing mocks for extract_mcp_auth_context and set_auth_context,
causing the handler to fail silently in the except block instead of reaching
session_manager.handle_request. This mirrors the fix already applied to the
sibling test_sse_mcp_handler_mock.
Co-authored-by: Ishaan Jaff <ishaan-jaff@users.noreply.github.com>
* fix(ci): route OpenAI models through chat completions in pass-through tests
The test_anthropic_messages_openai_model_streaming_cost_injection test fails
because the OpenAI Responses API returns 400 for requests routed through the
Anthropic Messages endpoint. Setting LITELLM_USE_CHAT_COMPLETIONS_URL_FOR_ANTHROPIC_MESSAGES=true
routes OpenAI models through the stable chat completions path instead.
Cost injection still works since it happens at the proxy level.
Co-authored-by: Ishaan Jaff <ishaan-jaff@users.noreply.github.com>
* fix(ci): fix assemblyai custom auth and router wildcard test flakiness
1. custom_auth_basic.py: Add user_role='proxy_admin' so the custom auth
user can access management endpoints like /key/generate. The test
test_assemblyai_transcribe_with_non_admin_key was hidden behind an
earlier -x failure and was never reached before.
2. test_router_utils.py: Add flaky(retries=3) and increase sleep from 1s
to 2s for test_router_get_model_group_usage_wildcard_routes. The async
callback needs time to write usage to cache, and 1s is insufficient on
slower CI hardware.
Co-authored-by: Ishaan Jaff <ishaan-jaff@users.noreply.github.com>
* ci: retrigger CI pipeline
Co-authored-by: Ishaan Jaff <ishaan-jaff@users.noreply.github.com>
* fix(mypy): use LitellmUserRoles enum instead of raw string in custom_auth_basic
Fixes mypy error: Argument 'user_role' has incompatible type 'str'; expected 'LitellmUserRoles | None'
Co-authored-by: Ishaan Jaff <ishaan-jaff@users.noreply.github.com>
* fix: don't close HTTP/SDK clients on LLMClientCache eviction (#22926)
* fix: don't close HTTP/SDK clients on LLMClientCache eviction
Removing the _remove_key override that eagerly called aclose()/close()
on evicted clients. Evicted clients may still be held by in-flight
streaming requests; closing them causes:
RuntimeError: Cannot send a request, as the client has been closed.
This is a regression from commit fb72979432. Clients that are no longer
referenced will be garbage-collected naturally. Explicit shutdown cleanup
happens via close_litellm_async_clients().
Fixes production crashes after the 1-hour cache TTL expires.
* test: update LLMClientCache unit tests for no-close-on-eviction behavior
Flip the assertions: evicted clients must NOT be closed. Replace
test_remove_key_closes_async_client → test_remove_key_does_not_close_async_client
and equivalents for sync/eviction paths.
Add test_remove_key_removes_plain_values for non-client cache entries.
Remove test_background_tasks_cleaned_up_after_completion (no more _background_tasks).
Remove test_remove_key_no_event_loop variant that depended on old behavior.
* test: add e2e tests for OpenAI SDK client surviving cache eviction
Add two new e2e tests using real AsyncOpenAI clients:
- test_evicted_openai_sdk_client_stays_usable: verifies size-based eviction
doesn't close the client
- test_ttl_expired_openai_sdk_client_stays_usable: verifies TTL expiry
eviction doesn't close the client
Both tests sleep after eviction so any create_task()-based close would
have time to run, making the regression detectable.
Also expand the module docstring to explain why the sleep is required.
* docs(AGENTS.md): add rule — never close HTTP/SDK clients on cache eviction
* docs(CLAUDE.md): add HTTP client cache safety guideline
* [Fix] Install bsdmainutils for column command in security scans
The security_scans.sh script uses `column` to format vulnerability
output, but the package wasn't installed in the CI environment.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: handle string callback values in prometheus multiproc setup
When callbacks are configured as a plain string (e.g., `callbacks: "my_callback"`)
instead of a list, the proxy crashes on startup with:
TypeError: can only concatenate str (not "list") to str
Normalize each callback setting to a list before concatenating.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* bump: version 1.82.2 → 1.82.3
* fix(test): update test_startup_fails_when_db_setup_fails for opt-in enforcement
The --enforce_prisma_migration_check flag is now required to trigger
sys.exit(1) on DB migration failure, after #23675 flipped the default
behavior to warn-and-continue.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix(cost_calculator): use model name for per-request custom pricing when router_model_id has no pricing
When custom pricing is passed as per-request kwargs (input_cost_per_token/output_cost_per_token),
completion() registers pricing under the model name, but _select_model_name_for_cost_calc was
selecting the router deployment hash (which has no pricing data), causing response_cost to be 0.0.
Now checks whether the router_model_id entry actually has pricing before preferring it.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: Ishaan Jaff <ishaan-jaff@users.noreply.github.com>
Co-authored-by: Ishaan Jaff <ishaanjaffer0324@gmail.com>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
* feat(proxy/utils.py): track pre-call hooks in OTEL
some pre call hooks can cause latency in high traffic - make sure this is tracked
* fix(router.py): move redis call on deployment_callback_on_success to pipeline operation
reduces p99 latency by half when redis is enabled
* fix(parallel_request_limiter_v3.py): only run check if any item has rate limits set
Prevents unnecessary latency added by rate limit checks
* test: add unit tests
* Latency Improvements: only track tpm/rpm usage when set on deployment+ LLM Caching - use an in-memory cache to reduce redis calls + OTEL - track time spent on LLM caching (#13472)
* fix(router.py): only track usage for deployments with tpm/rpm set
ensures additional latency avoided for non-tpm/rpm models
* fix(caching_handler.py): log time spent on request get cache to OTEL
enables easy debugging of call latency
* fix(caching_handler.py): use dual cache object for in-memory caching + trace redis call within caching handler
* fix(caching_handler.py): working in-memory cache for redis calls
ensures dual cache works when redis cache setup for llm calls
makes calls quicker by only checking redis when in-memory cache missed for llm api call
* test: remove redundant test
* test: add unit tests
* fix unsupported operand type(s) for +=: 'NoneType' and 'str' on clientside auth creds for responses
* fix the client side auth to use correct metadata
* add more tests
* fix tests
* fix(cost_calculator.py): handle custom pricing at deployment level for router
* test: add unit tests
* fix(router.py): show custom pricing on UI
check correct model str
* fix: fix linting error
* docs(custom_pricing.md): clarify custom pricing for proxy
Fixes https://github.com/BerriAI/litellm/issues/8573#issuecomment-2790420740
* test: update code qa test
* fix: cleanup traceback
* fix: handle litellm param custom pricing
* test: update test
* fix(cost_calculator.py): add router model id to list of potential model names
* fix(cost_calculator.py): fix router model id check
* fix: router.py - maintain older model registry approach
* fix: fix ruff check
* fix(router.py): router get deployment info
add custom values to mapped dict
* test: update test
* fix(utils.py): update only if value is non-null
* test: add unit test
* fix(route_llm_request.py): move to using common router, even for client-side credentials
ensures fallbacks / cooldown logic still works
* test(test_route_llm_request.py): add unit test for route request
* feat(router.py): generate unique model id when clientside credential passed in
Prevents cooldowns for api key 1 from impacting api key 2
* test(test_router.py): update testing to ensure original litellm params not mutated
* fix(router.py): upsert clientside call into llm router model list
enables cooldown logic to work accurately
* fix: fix linting error
* test(test_router_utils.py): add direct test for new util on router
* docs(sidebar.js): docs for support model access groups for wildcard routes
* feat(key_management_endpoints.py): add check if user is premium_user when adding model access group for wildcard route
* refactor(docs/): make control model access a root-level doc in proxy sidebar
easier to discover how to control model access on litellm
* docs: more cleanup
* feat(fireworks_ai/): add document inlining support
Enables user to call non-vision models with images/pdfs/etc.
* test(test_fireworks_ai_translation.py): add unit testing for fireworks ai transform inline helper util
* docs(docs/): add document inlining details to fireworks ai docs
* feat(fireworks_ai/): allow user to dynamically disable auto add transform inline
allows client-side disabling of this feature for proxy users
* feat(fireworks_ai/): return 'supports_vision' and 'supports_pdf_input' true on all fireworks ai models
now true as fireworks ai supports document inlining
* test: fix tests
* fix(router.py): add unit testing for _is_model_access_group_for_wildcard_route
* fix(main.py): support 'mock_timeout=true' param
allows mock requests on proxy to have a time delay, for testing
* fix(main.py): ensure mock timeouts raise litellm.Timeout error
triggers retry/fallbacks
* fix: fix fallback + mock timeout testing
* fix(router.py): always return remaining tpm/rpm limits, if limits are known
allows for rate limit headers to be guaranteed
* docs(timeout.md): add docs on mock timeout = true
* fix(main.py): fix linting errors
* test: fix test
* docs(config_settings.md): document all router_settings
* ci(config.yml): add router_settings doc test to ci/cd
* test: debug test on ci/cd
* test: debug ci/cd test
* test: fix test
* fix(team_endpoints.py): skip invalid team object. don't fail `/team/list` call
Causes downstream errors if ui just fails to load team list
* test(base_llm_unit_tests.py): add 'response_format={"type": "text"}' test to base_llm_unit_tests
adds complete coverage for all 'response_format' values to ci/cd
* feat(router.py): support wildcard routes in `get_router_model_info()`
Addresses https://github.com/BerriAI/litellm/issues/6914
* build(model_prices_and_context_window.json): add tpm/rpm limits for all gemini models
Allows for ratelimit tracking for gemini models even with wildcard routing enabled
Addresses https://github.com/BerriAI/litellm/issues/6914
* feat(router.py): add tpm/rpm tracking on success/failure to global_router
Addresses https://github.com/BerriAI/litellm/issues/6914
* feat(router.py): support wildcard routes on router.get_model_group_usage()
* fix(router.py): fix linting error
* fix(router.py): implement get_remaining_tokens_and_requests
Addresses https://github.com/BerriAI/litellm/issues/6914
* fix(router.py): fix linting errors
* test: fix test
* test: fix tests
* docs(config_settings.md): add missing dd env vars to docs
* fix(router.py): check if hidden params is dict
* fix(ollama.py): fix get model info request
Fixes https://github.com/BerriAI/litellm/issues/6703
* feat(anthropic/chat/transformation.py): support passing user id to anthropic via openai 'user' param
* docs(anthropic.md): document all supported openai params for anthropic
* test: fix tests
* fix: fix tests
* feat(jina_ai/): add rerank support
Closes https://github.com/BerriAI/litellm/issues/6691
* test: handle service unavailable error
* fix(handler.py): refactor together ai rerank call
* test: update test to handle overloaded error
* test: fix test
* Litellm router trace (#6742)
* feat(router.py): add trace_id to parent functions - allows tracking retry/fallbacks
* feat(router.py): log trace id across retry/fallback logic
allows grouping llm logs for the same request
* test: fix tests
* fix: fix test
* fix(transformation.py): only set non-none stop_sequences
* Litellm router disable fallbacks (#6743)
* bump: version 1.52.6 → 1.52.7
* feat(router.py): enable dynamically disabling fallbacks
Allows for enabling/disabling fallbacks per key
* feat(litellm_pre_call_utils.py): support setting 'disable_fallbacks' on litellm key
* test: fix test
* fix(exception_mapping_utils.py): map 'model is overloaded' to internal server error
* test: handle gemini error
* test: fix test
* fix: new run
