Wrap the context-management editors, the end-of-stream chunk builder,
acount_tokens, the compression interception hook, the passthrough
interrupted-stream recovery, the A2A usage counters, and the semantic
cache embedding truncation in asyncify so a multi-megabyte payload no
longer stalls the worker's event loop while it is tokenized
The pass-through suite now drains the process-global logging worker
from an autouse conftest fixture so work queued on one test's loop
cannot fire against the next test's callbacks
Resolves LIT-7190
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
The failure hook's usage estimate and the project ITPM reservation both
called litellm.token_counter inline on the event loop, so a large request
that failed or hit the limiter stalled the gateway the same way the
count_tokens endpoints did. Both now run through asyncify. The loop-lag
probe the existing tests used moves into a shared helper that warms the
tokenizer first, and two new tests fail when either count runs inline
A string above TOKEN_COUNTER_MAX_EXACT_CHARS was counted from its first cap characters and scaled, so a string whose start tokenizes unlike its end got a skewed count, and that count reaches fallback billing when the provider sends no usage. The estimate now tokenizes 16 evenly spaced samples that together total the cap and scales their sum by the string's length, keeping the same bound on work while tracking the whole string
Both proxy token counting endpoints already count in a worker thread, but the
HuggingFace tokenizer's encode holds the GIL for the whole call, so a 600k-token
count on a Claude model still froze the event loop for up to 0.8 s and every
other request with it. Count through encode_batch_fast, which releases the GIL,
and tokenize at most TOKEN_COUNTER_MAX_EXACT_CHARS characters of any one string
(default 4,000,000), scaling the exact count of that prefix by the string's
length above it so the largest payloads stay bounded.
* fix(bedrock): keep x-amzn-RequestId on chat error responses
Bedrock chat error paths built BedrockError from only a status code and a
message, so the provider response headers were gone before exception mapping
ran and the proxy had nothing to forward. AWS support needs x-amzn-RequestId
to investigate a server-side error.
- converse and invoke chat handlers pass the real headers and response when
they turn an httpx.HTTPStatusError into a BedrockError, and read the body
through error_response_text so a streamed body nobody read does not throw
- every bedrock chat get_error_class honors the headers it is already handed:
invoke, moonshot, bedrock-hosted openai, agentcore and the invoke agent
- BedrockError carries those headers into the response it synthesizes when a
caller has headers but no response, skipping values httpx cannot carry
- the bedrock 500 mapping forwards the provider response like its 4xx and 503
siblings instead of fabricating a blank one
The proxy now returns llm_provider-x-amzn-requestid on Bedrock chat errors.
* fix(bedrock): keep request-id on text-classified errors
The context-window and image branches of _map_bedrock_exception built their
litellm exception without the provider response, so a Bedrock 400 classified
by its body text lost x-amzn-RequestId while the sibling branches kept it.
Also narrows the new BedrockError types and trims its docstrings.
* chore(bedrock): drop the docstrings on the new error helpers
* fix(bedrock): keep request-id on every error path that has one
The ticket's root cause is that every BedrockError raise site under
litellm/llms/bedrock/ was built from status and message alone. The first
commits covered the chat and invoke handlers; this covers the rest.
Embeddings, rerank, image generation, image edit, count tokens, search and
the transformation layers now hand on the provider response or its headers,
and both bedrock_mantle configs return a BedrockError instead of the
OpenAI error that drops them.
Two blockers surfaced while verifying the streaming path. The trailing
`except Exception` in make_call and make_sync_call swallowed the BedrockError
raised a few lines above, relabelling a provider status as a 500, and the
non-200 branch read an unread streamed body, which throws.
The raise sites left alone have no provider response to carry: timeouts,
credential and config errors, and mid-stream event frames.
* fix(bedrock): forward provider headers from the count tokens route
The count tokens route converts BedrockError into an HTTPException, and dropped
the headers the handler had just kept, so that route still lost the request id.
get_response_headers now takes a Mapping so an httpx.Headers can be handed to it
without a copy.
* fix(bedrock): classify every bedrock surface through BedrockError
Eleven bedrock configs still inherited a provider-agnostic get_error_class
that builds a blank response, so the request id was gone before the proxy
read it. Claude platform, bedrock anthropic-messages, both image edit
configs, passthrough, realtime, vector stores and agentcore search now
return BedrockError, and a parametrized audit drives all 36 configs.
* fix(proxy): keep provider headers on the httpx status error branch
_handle_llm_api_exception forwards safe_headers on every branch except the
httpx.HTTPStatusError one, which the bedrock passthrough route reaches, so
the request id was dropped before the client saw the response.
* fix(bedrock): keep the request id on the timeout mappings
Timeout takes no response argument, so the three bedrock timeout branches
dropped the provider headers even when the upstream answered 408 or 504
with an x-amzn-RequestId. They now ride on the exception, already
llm_provider-prefixed, which is the form the proxy emits.
* fix(bedrock): keep the provider response on mapped timeouts
The previous round attached llm_provider-prefixed headers directly to the
Timeout. That shadowed the raw upstream headers for _get_response_headers,
so router cooldown and fallback cooldown stopped honouring retry-after on
bedrock 408/504 replies.
Give Timeout an optional response instead, the way every other mapped
bedrock exception already carries one. Retry logic reads the raw
retry-after off the response, and the proxy prefixes those headers on the
way out, so clients still see llm_provider-x-amzn-requestid.
* chore(bedrock): drop the explanatory comment on Timeout.response
Take staging's test_bedrock_knowledgebase_hook.py, which drops the duplicate
embedding_executor parameter that turned the lint check red, and make the two
cross-module helpers this branch added public (raise_denied_scoped_mcp_access
and routes_through_gateway) so the private-usage budget stays at its base count
Encrypt secret maps at the shared persistence boundary, preserve plaintext API/runtime views, and extend rotation and migration scanning to legacy rows.
Co-authored-by: Claude Code <noreply@anthropic.com>
Move the only guidance worth keeping into the parity assertion message so a
failing run tells the contributor to add a catalog entry instead of growing
the frozen unlisted set.
Claude-Session: https://claude.ai/code/session_011Tn3657NkV6ojLqewL64Kb
The Add Model provider dropdown is driven entirely by provider_create_fields.json,
and chatgpt had no entry there, so the documented ChatGPT subscription setup was
unreachable from the Admin UI. Add the entry plus the dashboard enum, slug, logo and
placeholder mappings so the provider can be selected and its cost-map models listed.
The entry carries no credential fields on purpose: the chatgpt backend ignores
api_key and api_base and signs in through the device-code auth file on the proxy
host, so any field here would be inert.
Add a parity test that every LlmProviders value is either listed for Add Model or
frozen in an explicit unlisted set, so a new backend provider cannot silently miss
the dropdown again.
Claude-Session: https://claude.ai/code/session_011Tn3657NkV6ojLqewL64Kb
* fix(guardrails): accept on_violation block and alert for mcp_security
The MCP Security policy template sends on_violation: "block", but the shared
LitellmParams model only allowed the /v1/realtime values "warn" and
"end_session", so POST /guardrails returned 422 before the MCP guardrail was
initialized. Widen the literal to include the MCP actions, map every non-alert
value to MCP's default "block" at init, and regenerate the lazy OpenAPI
snapshot and dashboard API types
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(guardrails): restrict on_violation block/alert to mcp_security and keep legacy MCP mapping
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(guardrails): return 422 when PATCH sets an mcp_security-only on_violation on another guardrail
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---------
Co-authored-by: yassin <yassin@berri.ai>
Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(router): keep provider response headers on streaming chat completions
The Router re-wraps a deployment's CustomStreamWrapper in FallbackStreamWrapper
(and its sync twin) so a mid-stream failure can fail over. Neither wrapper
forwarded `_response_headers`, so every streaming chat completion handed the
proxy's callbacks and its response-header builder a wrapper with no provider
headers, and a successful mid-stream fallback still published the failed
deployment's identity, `x-request-id` and rate limit counters.
Forward `_response_headers` into both wrappers, repoint the wrapper at the
deployment that served the stream once a fallback takes over, and rebuild the
proxy's response headers from that deployment while `create_response` still has
the first chunk buffered.
* fix(router): follow a nested fallback to the deployment that served the stream
A fallback the router picks is itself a fallback-aware wrapper, and it only
repoints at its own fallback once it yields, so reading its hidden params at
selection time named a deployment that produced no output. Re-read them when
the first fallback item arrives, which is still before the proxy commits
response headers.
Also addresses review feedback: the streaming header builder reads self.data
instead of taking a coarse request_data parameter, and the new test recorder
local is Final.
* test(router): cover the fallback header adoption helper directly
The router_code_coverage gate wants every router.py function named in a
router test, and this also pins the weak-reference behavior: a wrapper
collected mid-stream must not break the generator still draining it.
* refactor(proxy): take a read-only mapping for the model-id lookup
_get_model_id_from_response only reads its request payload, so a Mapping
says what it needs and the two metadata hops are narrowed instead of
assumed to be dicts.
* test: drop mutable recorder locals and routine comments from the new tests
An AsyncMock await_count and an asyncio.Event say the same thing as a
list and a dict that the test mutates.
* chore(router): justify the two rebinds in the fallback loops
Both are the one-shot re-read that follows a nested fallback, so they get
the repo's rebind-ok note like the rest of the file.
* feat(skills): semantic search over the LiteLLM-hosted skill registry
Adds GET /v1/skills?query= (custom_llm_provider=litellm_proxy) and a
skill_search MCP virtual tool, ranking the caller's accessible skills by
semantic similarity, mirroring the A2A agent registry search (LIT-6309).
Also fixes a pre-existing bug where create_skill() dropped description and
instructions for the litellm_proxy provider, which left every LiteLLM-hosted
skill with no searchable text.
* fix(mcp): coerce skill_search top_k instead of raising 500 on malformed input
The MCP-REST skill_search dispatch validated raw tool arguments through a
pydantic model directly, so a non-numeric top_k raised a ValidationError
that the endpoint's catch-all turned into an HTTP 500. Mirrors the
agent_search branch's tolerant coerce_top_k handling instead.
* fix(skills): enforce key limits on search embeddings and bound the semantic index
Semantic search embeddings now run the same pre_call_hook the /embeddings
route runs, so key rate limits, budgets and guardrails apply before the
embedding model is called. The shared SemanticTextIndex caps cached vectors
and evicts the least recently searched entries, and each skill's embedded
text is capped so one skill cannot inflate the embedding batch
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(skills): surface proxy 429s from search embeddings instead of a 503
ProxyRateLimitError is also an OpenAIError, so the search engine was folding
a key rate limit into skill_search_unavailable. Proxy HTTPExceptions now
propagate so the caller gets the same 429 the /embeddings route returns
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(skills): import assert_never from typing_extensions for Python 3.10
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(skills): embed the request as the pre-call hooks returned it, not the original text
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* refactor(skills): keep the litellm_proxy provider check for GET /v1/skills?query= inside llms/
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* test(skills): move the GET /v1/skills?query= endpoint tests under tests/test_litellm/proxy
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---------
Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* feat(ui): add key-scoped auto-router usage tab
GET /auto_router/benchmarks takes an optional api_key filter, applied in the
rollup aggregate on the primary key's leading column. Proxy admins get a
separate Auto-router usage tab on key detail pages with spend, baseline,
savings, tier routing, cache metrics and the existing router selector
* fix(ui): share key analytics date range
* feat(proxy): resolve root_path per request from SERVER_ROOT_PATHS
One deployment can encode exactly one client-visible URL path prefix
today: SERVER_ROOT_PATH is a scalar stamped onto the app at startup, so
a pod fronting several ingress prefixes 404s every prefix but one before
any handler runs, and MCP OAuth discovery can emit only one prefix's
URLs (RFC 9728 section 3 exact-match fails for the rest).
Add an opt-in outermost ASGI middleware that matches the request path
against a configured prefix list (SERVER_ROOT_PATHS, comma-separated) on
a segment boundary and sets scope["root_path"] for that request only.
Everything downstream is stock Starlette: route matching strips
root_path so routes stay registered root-relative, and request.base_url
re-includes it, so the discovery documents' resource and the 401
challenges' resource_metadata land under the prefix the client actually
called — with no discovery-builder changes.
LazyFeatureMiddleware now strips the scope root_path (falling back to
the cached SERVER_ROOT_PATH scalar) before feature prefix matching, so
lazily-registered routers — the MCP OAuth discovery router among them —
load under per-request prefixes.
Follow-up to the routing discussion on #35226; composes with, but does
not depend on, #35576.
* fix(proxy): import Sequence from collections.abc (ruff UP035 strict-budget gate)
* review(greptile): trim implementation commentary; fixture-own MCP registry state in tests
Addresses both P2s from the first Greptile pass:
- per_request_root_path_middleware.py (and the related _lazy_features /
proxy_server comments) cut down to the constraints the code cannot
express, per repo comment guidance
- the new discovery tests no longer clear/repopulate the shared MCP
registry inline; a fixture snapshots it, hands the test an empty
registry, and restores it afterwards so no state leaks between cases
* fix(lint): mutable-ok marker on the prefix accumulator (LIT002 type-discipline gate)
* fix(proxy): tie 401 challenges and get_custom_url to the per-request root_path
The per-request root_path middleware sets scope["root_path"] to the
prefix the client actually called, but the OAuth 401 challenges
(raise_user_oauth_challenge / raise_token_exchange_challenge) still
built their resource_metadata from SERVER_ROOT_PATH. On a pod fronting
several prefixes, the challenge advertised a discovery URL under a
different prefix than the discovery document served — the two
disagreed on where the resource metadata lives, and a strict RFC 9728
client refused the challenge. Route the challenges through a small
ContextVar the middleware populates so they read the same effective
root_path Starlette resolves the request under.
The same accessor fixes get_custom_url: when a request lives under a
SERVER_ROOT_PATHS-matched prefix, request.base_url already carries it,
so appending the SERVER_ROOT_PATH scalar on top produced e.g.
/tenant-a/legacy/sso/callback — a path that does not exist. Reading
the per-request prefix instead (and relying on join_paths's tail-dedup)
keeps SSO login/callback URLs under one prefix — the one the request
actually arrived on.
Fallback: outside a request (module-load-time UI URL builders,
background tasks) the ContextVar is unset and the accessor reads
SERVER_ROOT_PATH, matching get_server_root_path() so scalar-only
deployments are byte-identical.
* fix(mcp): challenge URL under per-request prefix must route, and mock parity
Two follow-ups to the review fix that made the 401 challenge use the
per-request root_path:
1. oauth_protected_resource_path must pick the URL structure that
actually routes for the mechanism in use:
- The scalar SERVER_ROOT_PATH deployment registers the well-known
routes with the prefix INSERTED (via well_known_root_suffix at
import time), matching RFC 8414 §3. The challenge URL must use the
same insertion or a client fetching it 404s.
- The per-request SERVER_ROOT_PATHS deployment can't register routes
per prefix; PerRequestRootPathMiddleware strips the prefix from
scope["path"] and the router matches the un-inserted route. The
URL must place the prefix BEFORE .well-known so the strip leaves a
matching path.
The previous fix used the insertion form for both, which 404'd the
discovery fetch on the per-request path — the discovery doc and the
challenge would then disagree on where the resource metadata lives,
the very failure the review flagged. End-to-end verified: the URL
the challenge advertises routes and the doc's `resource` field
equals the URL the client originally called (RFC 9728 §3).
2. get_request_root_path now delegates its fallback through
get_server_root_path() instead of reading the env directly, so every
existing `monkeypatch.setattr("litellm.proxy.utils.get_server_root_path"`
test override keeps working. This unstubbed the mock on the /v2/login
test that failed on the last CI run.
Plus the lint budget: annotate the local accumulator Final, tag the
scope["root_path"] rewrite as an intentional ASGI-contract mutation,
tag the reused `path`/`root_path` rebinds in LazyFeatureMiddleware, and
add reason strings to the two new PLC0415 lazy-import noqas.
* test(mcp): pin the reviewer's expected end-state — challenge URL routes, resource matches called URL
End-to-end regression test that mounts the discoverable router + the
per-request root_path middleware, hits an MCP endpoint that raises
raise_user_oauth_challenge, fetches the resource_metadata URL the
challenge advertises, and checks the returned document's `resource`
equals the URL the client originally called (RFC 9728 §3 exact match).
Covers /tenant-a, /tenant-b, and the unprefixed path on the same app so
a regression on any prefix — challenge URL 404s, or doc emits a
different prefix than the client called — fails at this test rather
than in a strict MCP client's discovery.
---------
Co-authored-by: gym-cmd <186399764+gym-cmd@users.noreply.github.com>
* fix(spend_logs): keep partition DDL transactions alive for their statement timeout
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* style: ruff format changed files
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(lint): avoid dict-literal kwargs and keep cast-ok on the cast line
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* fix(lint): cast at the call site instead of widening PrismaClient.tx
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* test(spend_logs): require partition tx timeout to strictly exceed statement bound
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---------
Co-authored-by: jesus <jesus@berri.ai>
Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* test(e2e): prove the virtual key lifecycle on every replica
Walks one virtual key through create, read, partial update, clear, enforce
and delete against a live proxy and database, reading every write back on
every gateway replica.
The management suite already had single write-then-read tests for keys, but
none of them proved that a partial /key/update leaves the untouched fields
alone, that an explicit null clears a field, or that a write is visible on
more than the one gateway that took it.
Adds read_back_everywhere to the shared ProxyClient: it polls a GET path on
every URL in PROXY_REPLICA_URLS until each replica's parsed body satisfies
the caller's predicate, and fails naming the replica that never converged.
The CLEAR sentinel in the e2e models makes an explicit JSON null expressible
in a body the transport otherwise strips of None fields.
Documents /key/update's merge patch semantics on the endpoint docstring.
* test(e2e): prove key revocation and field preservation on every replica
Applies the findings from an adversarial review of the first commit.
The delete step only checked that chat was refused on the gateway that took
the write, so it would have passed while a sibling gateway kept serving the
deleted key. It now serves one call from every replica first, so each has the
key cached and the delete has something to revoke everywhere, then polls every
replica for the refusal.
The file also carried its own poll loop that tested the deadline before
attempting, so it gave up one attempt early and skipped the attempt landing
exactly on the deadline. It now shares the harness helper, which is generic
over the polled value rather than over a parsed body, so the same loop covers
both the info read-back and the chat refusal.
The model the enforcement step registers now carries a unique marker in its
alias, matching every other deployment this suite creates, so concurrent runs
never share one model group.
The docstring sentence claimed an explicit null clears any field. It does not:
the metadata-backed fields merge into stored metadata, where a null is a silent
no-op, and only the key's own columns clear. Regenerating the dashboard types
picks up the corrected text.
* fix(e2e): delete a deployment that never becomes servable
Registering a model posts /model/new and then waits for every replica to list
it. When that wait timed out the deployment already existed in the database but
its id had never been returned, so no caller could delete it and the row
outlived the run. It is now deleted before the failure propagates.
Found by review on the key lifecycle suite, whose module fixture registers a
deployment this way, but every caller of the shared helper had the same
exposure.
* docs(e2e): drop the duplicated notes from the lifecycle docstrings
The delete method restated what the warm-up helper already explains, and the
module restated the merge patch rule that the endpoint and the request model
both document.
* test(e2e/ui): cover member role and budget edits, member permission delegation, and team guardrail removal
Three Playwright specs for the Teams flows enterprise customers hit most, each
owning its fixtures and proving the mutation through a read-back rather than a
toast.
- teamMemberEdit: an admin edits a member's team role and per-member budget,
and both survive a reload of the Members table
- memberPermissions: a plain member is refused /key/generate for their team,
a team admin grants it on the Member Permissions tab, and the member then
creates a team key that serves a real completion
- teamGuardrailRemoval: clearing a team's only guardrail on the Settings tab
really clears it, and traffic the guardrail refused starts serving again
* test(e2e/ui): make the new team specs safe to run in parallel
Fixture ids came from Date.now(), so two repeats starting in the same
millisecond minted the same user id: one got a 409 and the loser's teardown
deleted the user the other was still signed in as. Ids now carry a random
suffix.
Also move the member-permissions setup inside the cleanup-protected block so a
half-finished setup cannot leak a team, and close both browser contexts the
test opens.
