Offload the per-block hashing through offload_token_count on both the pre-call
read and the success-event write, hash raw bytes as base64 instead of raising,
drop the unused serialize_object helper, and bind the chained digest, the
message envelope, and the bytes path in the regression tests
A stream counted before its usage is known increments TPM by zero, so the
worker that served it never refreshed its local TPM value from Redis and
the first byte headers reported the token count another worker had already
consumed. Both pipeline operations now always run, matching the pre-change
callback, so the returned values refresh both worker local keys
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Router.make_call now increments the deployment TPM/RPM counter before set_response_headers reads remaining usage, so the headers carry post-increment values directly and the in-flight subtraction workaround from LIT-2719 is removed. deployment_callback_on_success adds only the tokens not yet counted (streams) and never a second request. The counter key uses the resolved deployment name so wildcard routes are read back correctly, and the proxy strips the router-owned counted-tokens marker from client metadata
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
_handle_clientside_credential registered the per-request Deployment it built for
a client-supplied api_key/api_base via upsert_deployment, which added it to
self.model_list under the shared model_name. That made a request-scoped
credential a permanent, load-balanced deployment that any later caller of the
same model group could be routed onto, reaching the provider with someone
else's forwarded credential.
The per-request Deployment still gets its own stable id for cooldown and
logging identity; it is just never registered with the router.
Resolves LIT-7811
The router clamps a negative request_retry_count found in request metadata before counting a failure, and the proxy strips a client-supplied request_retry_count with the other router-reserved metadata fields. The rust OCR lifecycle test that trips the per-request cap now plants request_retry_count instead of attempted_retries, which the cap no longer reads since the previous commit
num_retries_per_request has always capped the retries of one request with its fallback hops included. #40930 started reading the per-hop attempted_retries counter instead, and every fallback hop restarts that counter at zero, so a request could spend a fresh retry budget on each hop and the legacy fallback cap test started seeing the hop run.
Router.log_retry now also keeps request_retry_count on the request metadata, incremented on every retry and fallback hop and never truncated the way previous_models is, and max_retries_per_request_hit reads that count. The flat retry records, the litellm_metadata coverage and caps above four from #40930 stay as they are, and the legacy test goes back to its previous_models == 0 assertion.
Router.log_retry used to copy the failed attempt's kwargs and metadata into
metadata.previous_models. Nothing downstream read those copies, but they carried
client credentials into spend logs and grew the payload on every retry. Each
attempt now leaves a flat record (model group, deployment id, exception type and
string, attempt number), which drops RETRY_BREADCRUMB_EXCLUDED_KWARGS and the
per-retry credential masking.
num_retries_per_request was enforced from len(previous_models), which only
looked at the metadata bucket and never exceeded four records. The sync and
async client wrappers and the Rust lifecycle guard now read attempted_retries
from whichever metadata bucket the call carries.
Resolves LIT-7505
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Cooldown entries rode the router-wide DualCache, which re-reads a key that is
missing from memory at most once every 10s. A deployment benched on one replica
therefore kept taking traffic on its siblings for up to 10 seconds, and the same
shared in-memory tier could evict a live cooldown once 200 unrelated router keys
crowded it out, which sent even the benching replica back to the dead deployment.
CooldownCache now owns a DualCache over the router's Redis with a 1s read
interval and an in-memory tier that only holds cooldown keys. Redis is attached
lazily because the router builds the cooldown cache before it wires Redis up.
test_no_linear_scans_in_router: #39674 renamed heuristic_v2_router_limit_violation
to auto_router_capability_violation, so the allowlist entry stopped matching and the
same admin-only scan tripped the static check. Rename the entry to follow it.
tableScrolling.spec.ts: 9ba6cab889 (LIT-4738) gave the Tags and Model Hub tables
client-side pagination at 25 rows, so the 40 seeded rows no longer render on one
page. Select 50 rows per page before counting, as the Logs case already does.
test_no_linear_scans_in_router: #39468 added config_deployments() and
heuristic_v2_router_limit_violation(), which both scan the whole model_list
from admin-only paths (model add/upsert), so add them to the allowlist. The
allowlist becomes a mapping so each exemption carries its reason as data.
test_missing_model_parameter_curl: a request with no model is rejected by the
proxy when nothing can serve it and by the router when a wildcard or default
deployment exists, and by the upstream provider when a wildcard forwards it,
so the message text is not a stable contract. Assert the contract that holds
in every case: HTTP 400 with a non-empty error message.
test_model_group_info_e2e: /model_group/info resolves wildcards, so it can
never return "anthropic/*" verbatim. cc3f9cd65b rewrote the assertion to
expect the raw pattern after claude-3-5-haiku-20241022 left the price map,
which made it unsatisfiable. Assert the expansion instead.
test_should_derive_ocr_mapping_status_from_live_tests: the audit needs a
native bridge built with the trace-parity feature, which CI never builds, so
skip with the harness's own diagnostic instead of erroring. Extract that
check out of ensure_trace_bridge as trace_bridge_error so a pytest run
reports the state without kicking off a maturin rebuild.
The Router executor only routed a query embedding when the vector store
carried extra embedding configuration, so a store registered with no
embedding model at all always went to the Router and 500'd on the
s3_vectors default text-embedding-3-small when no deployment served it.
Route on whether the Router serves the model, which is the rule the
executor had before, and keep the request metadata on the SDK fallback so
the embedding stays attributed either way.
Milvus REST and Azure AI Search still embedded the query through the SDK, so
a bare Router alias as litellm_embedding_model kept failing after the executor
landed for Valkey. Both now share BaseQueryEmbeddingVectorStoreConfig, which
embeds through the injected executor, drops the empty litellm_embedding_config
requirement, and awaits aembedding on the async path.
The Router executor falls back to the SDK for models the Router does not
serve, so inline provider configs such as azure/text-embedding-3-large with
their own credentials keep working through the proxy.
Tests fake OpenAI and Milvus at the HTTP boundary with respx instead of
patching litellm.embedding.
* fix(search): forward search-tool params through the router, complete Parallel AI v1 param mapping
SearchAPIRouter dropped every parameter configured on a search tool, forwarding
only per-request kwargs. Any tool-level setting (mode, max_results, ...) was
silently lost on the way to the adapter, for every search provider.
Also completes the Parallel AI v1 search surface: after_date, fetch_policy,
location and include_domains now nest under advanced_settings instead of being
sent as unknown top-level fields, responses preserve search_id / session_id /
warnings / raw excerpts, and search cost is derived from the request mode and
the provider's reported usage rather than a single flat rate.
* fix(parallel_ai): stop a caller from pricing its own search request
`_parallel_ai_usage` carries the provider's reported usage into cost
calculation. It was only written when the response contained a usage block, so
a caller could pass `_parallel_ai_usage=[{"name": "sku_search", "count": 0}]`
and, whenever the provider omitted usage, bill $0.00 instead of $0.005 — the
value also reached the upstream request body as an unknown field.
The key is now stripped from inbound params and written unconditionally from
the parsed response, so only the provider can populate it.
* fix(parallel_ai): price fast search mode correctly
* test(parallel_ai): fake search at HTTP boundary
* fix(parallel_ai): tolerate null search result fields
---------
Co-authored-by: khushishelat <shelatkhushi@gmail.com>
* fix(router): resolve fallbacks against the tier a pre-routing hook selected
A complexity or auto router picks a tier behind the router group name, but
fallback lookup kept using kwargs["model"], which is still the router name. The
tier's configured chain never ran, so a provider failure on its first hop went
straight back to the client with "No fallback model group found for original
model_group=smart-router".
The hook assigns the selected model to a local only, and fallback resolution runs
on an outer kwargs dict that **kwargs already copied, so writing it there is not
visible. Record the selection in the metadata bucket instead, which is a nested
dict shared by reference across those copies and is how the router already
carries values back up, then key fallback lookup off it when present.
Applies to the generic, context-window, content-policy and weighted-failover
lookups. Reporting keeps using the router name, since that is what the caller
asked for.
Fixes#38832
* fix(router): annotate the recorded-selection helper with a read-only mapping
record_pre_routing_selection only reads the request kwargs, writing into the
nested metadata bucket it finds there, so Mapping states what it actually needs
and clears the LIT001 mutable-annotation budget without a suppression.
* test(router): assert the no-kwargs path leaks nothing
The tolerated-None case called the helper without checking anything, which the
test-quality gate counts as a test with no assertion. Assert that a fresh mapping
still reads back empty, so the case proves the call is a no-op rather than only
that it does not raise.
* fix(router): stop declaring loop-assigned locals Final in the selection helpers
Both helpers annotated a loop-assigned local as Final, which reassigns a Final on
every iteration and cost three basedpyright errors. Read the buckets through a
generator instead, so the write path iterates a for-target and the read path
resolves in one shot with next(), which also matches the functional style the
type-discipline rules ask for.
* style(router): apply ruff format to the selection helpers
* fix(router): derive the pre-routing tier fresh on every fallback hop
The metadata buckets also carry whatever the caller sent, so an inbound
pre_routing_selected_model let a client pick which fallback chain its
request fell into. A fallback hop also inherited the previous hop's tier,
so the second hop keyed its own failure off the tier that already failed
and never ran its own chain.
Clear the key at the top of async_function_with_fallbacks. Every hop
re-enters there, so only the hook that routed that hop can set it.
* fix(router): drop the cast at the fallback-hop clear call site
* feat(router): fall back on anthropic safeguard refusals on /v1/messages
---------
Co-authored-by: Priyansh Nandwana <nandwana.priyansh103@gmail.com>
Container create and list requests had no container ID to decode, so the
router called the provider handler directly and the OpenAI transformation
fell back to the global OPENAI_API_KEY. Proxies configured only with
model_list credentials sent Authorization: Bearer None. Route through
_ageneric_api_call_with_fallbacks when the caller passes a model, expose
the list endpoint's model query param to the router, and encode the
managed container ID on the async create path so follow-up calls route
to the same deployment.
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* 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
* test: unshadow the module handles the F811 sweep left behind, and pin the two live tests that went red with it
The F811 sweep in #37878 removed the fixture-local `import litellm` from four
conftests, but the bare `import litellm.proxy.proxy_server` a few lines below
still binds `litellm` as a function local, so `importlib.reload(litellm)` runs
before the name is assigned and every test in those directories errors at
setup. The `hasattr` guard on the line above already proves the module is
loaded, so the import only ever bound the name. Drop it, and enable F823 in
ruff-tests.toml, which flags all four sites at the failing line and would have
blocked the sweep
The same sweep renamed the `check_non_streaming_response` parameter but left
one read of `completion`, which now resolves to `litellm.completion`, and
removed an import whose side effect was the only thing making
`litellm.proxy.proxy_server` reachable in the moderation hook test. That test
already takes `monkeypatch`, so patch the router through it and stop leaking
the router into later tests
`test_content_policy_exception_openai` passed vacuously until #37887 turned it
into a real `pytest.raises`, and OpenAI no longer rejects a lyrics prompt with
a content policy error. Inject an AsyncOpenAI client whose transport answers
with OpenAI's own `content_policy_violation` rejection so the mapping to
ContentPolicyViolationError is exercised every run
`test_async_create_batch` hit a 409 cancelling a batch OpenAI had already
marked failed. The cancel step tolerated a completed batch but not a failed
one. Fold both guards into one helper that tolerates a failed batch only when
OpenAI's recorded error is the org's enqueued token limit, and prints the
batch's errors so the reason is in the log either way
* test: close the injected AsyncOpenAI client after the content policy test
* chore(lint): ratchet TQ005 down by the global mutation this branch cleared
* chore(lint): ratchet TQ005 to 2660 on the merged tree
* chore(lint): ratchet TQ005 to 2561 on the merged tree
* chore(lint): ratchet TQ005 to 2548 on the merged tree
A test that asserts on the error inside its own except block passes when the
call stops raising, because nothing runs the handler. That is the exact case
the test exists to catch, so the regression lands green.
Rewrites all 111 such blocks into pytest.raises, which fails when the call
succeeds, and selects PT017 in ruff-tests.toml so no new one lands.
A name bound twice keeps only the second binding. In `tests/` that is nearly
always a repeated import, harmless but misleading, and the same rule is what
catches the cases that are not harmless: a local that shadows an import the
module still calls, and a second `def test_x` that quietly replaces the first.
311 of the 344 sites were repeated imports and came out with ruff's own fix.
The remaining 33 needed a decision. Four modules imported a name they never
used because a local definition below already shadowed it. Two comprehensions
bound `call` over `unittest.mock.call`, which those modules import and use.
One test rebound the two module handles its nested reload closure had captured.
One class attribute shadowed an unused `status` import.
The load-test fixtures move to a conftest, which is how pytest is meant to share
them, so the test module no longer imports three fixture names it never calls.
The nine `prisma_client` parameters keep a narrow `noqa`: pytest resolves that
fixture by name before the body runs, so the parameter never shadows anything.
`pytest.raises(Exception)` with no `match=` passes on any error that broad. A
TypeError from a refactor, a botched fixture, an import that moved: all of them
read as the rejection the test claims to police, so the test goes green for the
wrong reason and stays green after the behaviour it guards is gone.
PT011 closes that gap for the 317 sites B017 could not reach, because B017 only
fires on a single-statement body with no `as e` binding. Each pattern here is the
message the code actually raised, recorded by running the sites under a plugin
that logged the concrete type and text per call site, so the assertions describe
observed behaviour rather than a guess. Where a site raises more than one message
across its parametrize cases, the pattern is an alternation of what was seen;
where the exception carries an empty `str()` and puts the text on `.message`, the
site keeps a narrow `noqa` with the reason.
PT014 removes four parametrize cases that were listed twice. The duplicate re-runs
an assertion that already passed, and it usually marks a case someone meant to
vary and forgot to edit.
* test: enforce PT012 so a pytest.raises block cannot hide dead assertions
`with pytest.raises(...)` stops at the first statement that raises. Anything
sequenced after it inside the block never runs, so an assertion written there is
never checked and the test still reports green.
Two sites were doing exactly that, and both assertions turned out to be wrong
once they started running. tests/llm_translation/test_prompt_factory.py asserted
the bedrock rejection names "requires at least one non-system message", which
holds. tests/proxy_unit_tests/test_proxy_server.py asserted the prisma startup
failure mentions "httpx.ConnectError", which never appears: the failure is an
httpx.ConnectError whose message is "All connection attempts failed", so that
test now asserts the type. Its DATABASE_URL override moves to monkeypatch, since
the old restore sat below the assertion and leaked the invalid URL into every
later DB test the moment the assertion started being able to fail.
The remaining 72 sites are rewritten without changing what they exercise: setup
that cannot raise moves above the block, a nested `patch` moves outside it, and
bodies with real control flow (a stream drain, an if/else on sync_mode, a
retry loop) move into a local closure the block calls.
Fixing PT012 unmasked two B017s, since ruff only reports a blind
pytest.raises(Exception) once the block holds a single statement.
tests/proxy_unit_tests/test_auth_checks.py narrows to the ProxyException
can_key_call_model actually raises. tests/local_testing/test_completion_cost.py
was asserting vertex_ai/medlm-medium has no cost entry, which stopped being true
at some point; that dead first half is gone and the rest of the test, which
checks medlm pricing resolves above zero, now runs instead of being skipped.
* chore(ci): ratchet TQ004 to 768 after the prisma test moved to monkeypatch
* feat(complexity_router): custom classifier plugins via classifier_type 'plugin'
Adds a third classification mode where an operator-supplied hook decides the
tier instead of the heuristic scorer or the LLM classifier. The hook implements
an async classify(context) returning a tier name (built-in value, tier_labels
label, or tier_definitions name) or None to decline; failures, timeouts, and
unknown tiers fall back exactly like a failed LLM classifier. The context
carries the request messages and metadata, including caller identity, so a
plugin can route by team, spend, or any business rule.
The plugin resolves from a dotted path at proxy startup with a load-time check
that classify is a coroutine function, and is closed off over HTTP like the
routing plugins list. Routing decisions record the new classifier_plugin cause.
tier_definitions now accepts classifier_type 'plugin' alongside 'llm'.
* fix(proxy): resolve plugin dotted paths in _delete_deployment before hashing ids
The db-sync reconcile re-reads the raw config and hashes litellm_params to
compute which ids the config wants served, but the router's ids were hashed
from the resolved params where plugin dotted paths are live instances. The
mismatched ids made the reconcile evict every plugin-bearing auto-router one
sync after startup, on any proxy with a database connected. This also affected
the existing routing plugins list, not just the new classifier plugin.
Resolving the plugins in _delete_deployment the same way load_config does makes
both sides hash the same canonical form. A plugin module broken on disk at
reconcile time skips cleanup instead of evicting valid deployments, matching
how a get_config failure is handled
* fix(complexity_router): treat non-string plugin verdicts as declines, centralize the empty-mapping sentinel
A hook returning a non-string raised inside resolve_classified_tier outside the
plugin exception boundary, failing the request instead of falling back. Also
moves the read-only empty mapping to constants.py per repo convention and moves
the classifier plugin product docs out of the package README for the docs repo
* refactor(complexity_router): rename the plugin classifier mode to classifier_type 'custom'
The mode value now names the operator's intent while classifier_plugin keeps
naming the mechanism; routing decisions keep the classifier_plugin cause
* refactor(proxy): pin plugin-bearing deployment ids from the raw params instead of resolving in the reconcile
Replaces the previous approach of re-running plugin resolution inside
_delete_deployment, which imported operator modules on every reconcile cycle
and skipped the whole cleanup pass when any one module was broken on disk.
load_config now stamps model_info.id from the raw litellm_params before
resolution swaps dotted paths for live instances, so the reconcile's raw-config
hash matches by construction and needs no resolution at all: a broken module
cannot stall cleanup for unrelated models, and any future param-transforming
resolution is covered by the same pin. _generate_model_id becomes a staticmethod
so the pin can run before the Router exists; its statically dead non-string key
branches are removed. Also documents candidate_models as an informational
snapshot for classifier plugins, unlike the narrowing surface RoutingPlugin
filters
* fix(router): restore _generate_model_id key handling, align classifier context with the routing-plugin pattern
The staticmethod conversion accidentally dropped the non-string-key branches
from _generate_model_id, a silent hash change for any params with non-string
keys; they are restored verbatim. The classifier plugin context now follows
the Router-level routing-plugin recipe exactly: structured messages come from
resolve_structured_messages over the raw messages, and the metadata key comes
from the shared get_metadata_variable_name_from_kwargs helper, which also
replaces the duplicated inline sniff in _pick_model_for_tier. This removes the
raw-or-resolved fallback where a plugin could silently receive resolved
messages when a call site forgot to pass the raw ones
* refactor(router): make generate_model_id public, guard classifier context construction
Two modules legitimately hash deployment ids with the same helper now (Router
and the proxy's config-load pin), so the private name was lying about its
audience and the cross-module call needed a pyright suppression; renaming it
public restores the static safety net. The classifier plugin's RoutingContext
construction moves inside the failure boundary, matching the LLM path where
litellm-side prompt building also falls back rather than failing the request,
and a prompt-only call with no message list is now covered by a test
Three groups, all verified by running the suite rather than by inspection.
18 files whose every test function carries an unconditional @pytest.mark.skip,
39 test functions in total. They are collected on every CI run and always skip,
so they advertise coverage the suite does not have. Reasons on the marks include
"AWS Suspended Account", "lakera deprecated their v1 endpoint" and "moved to
using 'otel' for logging"; 26 of the marks predate 2025.
30 test functions with a byte-identical body and identical decorators to a
sibling in the same file and class, differing only in name. Deleting one of each
pair removes no coverage. Four further candidates were excluded because they
override an inherited test, where deleting the override un-shadows the base
class implementation instead of removing a duplicate.
9 test functions that a later definition of the same name shadows, so Python
never binds them and pytest cannot collect them.
One file that is a demo script rather than a test; its own docstring says to run
it with python.
Verification: collecting the 26 edited files gives 2,492 node IDs before and
2,462 after. The 30 duplicate deletions account for exactly 30 removals, the 9
shadowed deletions account for 0 (confirming at runtime that they were never
collectable), nothing unexplained disappeared, and nothing new appeared. No
other test or module imports any deleted symbol.
LiteLLMCompletionStreamingIterator overrides __init__ without calling
super().__init__(), so completed_response was only set once the stream
reached RESPONSE_COMPLETED. On a mid-stream provider error the router's
_extract_partial_responses_usage read source_iterator.completed_response
during fallback recovery and raised AttributeError, masking the real
provider error (e.g. Anthropic 529) and bypassing configured retries and
fallbacks. Initialize the attribute to None so recovery degrades to no
partial usage instead of crashing.
Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
* feat(router): add per-deployment allowed_fails_policy and cooldown_time override support
Three bugs fixed in the router cooldown system: (1) deployment-level allowed_fails and
allowed_fails_policy in model_info now take precedence over router-level settings in
_should_cooldown_deployment; (2) failed fallback deployments now get evaluated for
cooldown via _trigger_cooldown_for_failed_deployment, bypassing the Logging dedup gate;
(3) DualCache promotes Redis cooldown entries using default 600s TTL instead of true
remaining cooldown time -- _corrected_active_cooldown now evicts expired entries and
corrects stale in-memory TTLs on backfill. Adds ServiceUnavailableError, BadGatewayError,
and NotFoundError fields to AllowedFailsPolicy and cooldown_time to LiteLLMParamsTypedDict.
* fix(router): gate fallback cooldown trigger on has_logged_async_failure; use only litellm_metadata for deployment ID
* fix(router): use X | Y union syntax to fix UP007 strict lint gate
* test(router_utils): add coverage for _trigger_cooldown_for_failed_deployment and has_logged_async_failure gate
* test(router_utils): cover deployment cooldown override and exception swallow paths
* fix(router): add InternalServerError/ServiceUnavailableError/BadGatewayError/NotFoundError to router-level get_allowed_fails_from_policy
* fix(router): format router.py and add router-level policy tests
* test(router): add CI-visible coverage for per-deployment cooldown policy
Tests for `_get_deployment_cooldown_policy`, `_resolve_allowed_fails_from_policy`,
and `_should_cooldown_based_on_deployment_policy` (cooldown_handlers.py), the
`_corrected_active_cooldown` branches in CooldownCache, and the four new
exception-type branches in `Router.get_allowed_fails_from_policy` (router.py) --
all in `tests/test_litellm/` which the enterprise-routing CI job runs.
* fix(router): use is not None guard for cooldown_time_override in should_cooldown_based_on_allowed_fails_policy
A cooldown_time_override of 0 was previously treated as falsy and silently
fell through to the router-level cooldown_time value. Switched to an explicit
is not None check so that zero is honored as a valid override.
Added a regression test covering the zero case.
* fix(router): honor has_logged_async_failure and metadata for fallback cooldown; support both model_info and litellm_params locations
Manual verification against a live proxy surfaced that the fallback-cooldown-gap
trigger never actually fired: the has_logged_async_failure check read a plain
attribute that Logging never sets (the real flag lives in model_call_details),
and the deployment_id lookup only trusted litellm_metadata, which regular chat
completions never populate (only batch/thread/file endpoints do). Router
overwrites model_info on whichever key is present before every attempt, so
metadata is equally authoritative there, not caller-controlled as previously
assumed. Also let allowed_fails/allowed_fails_policy/cooldown_time be set under
either model_info or litellm_params, each preferring its own canonical location.
* fix(router): fix ContentPolicyViolationError policy shadowing and partial-policy zero-threshold
Two bugs from Greptile review on PR #34416:
- ContentPolicyViolationError subclasses BadRequestError, so listing
BadRequestError first in _EXCEPTION_POLICY_FIELDS made the isinstance
check always match BadRequestError for content-policy errors, using the
wrong allowed_fails threshold. Reordered so the subclass is checked first.
- A deployment with a partial allowed_fails_policy and no deployment-wide
allowed_fails forced allowed_fails_override=0 for any exception type its
policy didn't cover, cooling the deployment down on the first unrelated
failure. Now defers to router-level behavior for uncovered exception
types instead of forcing an immediate cooldown.
* fix(router): only trust a metadata/litellm_metadata bucket the router itself wrote deployment info into
veria-ai flagged that preferring litellm_metadata whenever present could pick up a
caller-supplied litellm_metadata.model_info.id (preserved via allow_client_pricing_override)
instead of the metadata bucket the router actually populated for a regular completion's
fallback attempt, naming an arbitrary "victim" deployment for cooldown.
Router._update_kwargs_with_deployment() always writes model_info and
deployment_model_name into the same bucket together. Only trust a bucket that
carries deployment_model_name alongside model_info, since that marker is only
ever set by the router itself, not by request-body metadata.
* test(router): add regression coverage for ContentPolicyViolationError policy shadowing
The subclass-ordering fix in commit 38fe4e4490 had no regression test.
Verified the new test fails on the pre-fix ordering (asserts 2, got 10)
before restoring the fix, and confirmed the same behavior through the full
_should_cooldown_deployment call path against a real Router instance.
* fix(router): let explicit allowed_fails_policy entries override the generic 4XX cooldown exclusion
_is_cooldown_required skips cooldown evaluation for any 4XX status outside
{429, 401, 408, 404} by default, since a generic client error is usually not
the deployment's fault. BadRequestError and ContentPolicyViolationError both
carry status 400, so their AllowedFailsPolicy fields (BadRequestErrorAllowedFails,
ContentPolicyViolationErrorAllowedFails, both router-level pre-existing and the
new deployment-level ones) were silently unreachable: an operator could set
them to any value with no effect, since _is_cooldown_required blocked cooldown
evaluation before that policy was ever consulted.
_should_run_cooldown_logic now also checks whether an explicit allowed_fails_policy
entry (deployment-level or router-level) covers the exception's type, and if so,
proceeds with cooldown evaluation regardless of the generic status-code exclusion.
The exclusion remains the default for exception types with no explicit policy.
Verified live against a mock-triggered ContentPolicyViolationError (config-level
mock_response, azure/gpt-4.1-mini deployment) with BadRequestErrorAllowedFails=100
and ContentPolicyViolationErrorAllowedFails=0 on the same deployment: it now cools
down after exactly one ContentPolicyViolationError instead of never cooling down.
* fix(router): use the router-stamped failed_deployment_id for fallback cooldown targeting
Greptile flagged a real gap in the metadata-bucket-based deployment lookup:
for a generic-API-call fallback, the router writes the current attempt into
litellm_metadata, but a stale "metadata" bucket carrying the same
deployment_model_name marker (from an earlier point) would be picked first,
cooling the wrong deployment.
Router already has a more robust, pre-existing mechanism for this exact
problem: _set_failed_deployment_id_on_exception stamps the failing
deployment's id directly onto the exception at the point of failure,
immune to metadata-bucket ambiguity since a caller can't influence it and
it doesn't depend on which bucket the current call type happens to use.
It just wasn't called from _ageneric_api_call_with_fallbacks_helper's
except block, unlike _completion/_acompletion.
Added the missing call there (matching the existing pattern exactly), and
changed _trigger_cooldown_for_failed_deployment to prefer
exception.failed_deployment_id when present, falling back to metadata-bucket
inspection only for call paths that don't stamp it yet.
Verified live: the standard fallback-cooldown-gap scenario (two bad-key
deployments in a fallback chain) still correctly cools down both the
originally-called and fallback deployment.
* fix(router): address human review on per-deployment cooldown overrides
Scope allowed_fails_policy override to deployment-level only (a router-level
policy predates this feature and must keep its existing behavior), exempt
advisor-orchestration failures from the fallback cooldown trigger, keep the
single-deployment model group protection intact against a generic
deployment-level allowed_fails, make cooldown_time precedence consistent
across resolution paths, fix a falsy-zero swallowing bug in the router-level
allowed_fails fallback, and make allowed_fails_policy resolution fall through
to the next matching exception type instead of stopping at the first unset
field.
Also restrict allowed_fails/allowed_fails_policy/cooldown_time to model_info:
litellm_params gets copied into the actual provider request, so a router-only
setting placed there would leak into that request.
* test(router): update test_cooldown_handlers.py for the deployment-policy signature change
Surfaced by the rebase: this mirrored test file (tests/test_litellm/ mirrors
litellm/) predates the router_unit_tests/ coverage added earlier in this PR and
was still calling _should_cooldown_based_on_deployment_policy with its old
4-argument signature and asserting the now-removed litellm_params cooldown_time
location.
* test(router): update test_fallback_event_handlers.py for model_info-only cooldown_time
Another mirrored test file surfaced by the rebase that still asserted the
now-removed litellm_params.cooldown_time location.
* fix(router): match cooldown-duration precedence in the fallback path to the primary path
_trigger_cooldown_for_failed_deployment only checked deployment config before
falling back to the router default, skipping the response Retry-After header
step that Router.deployment_callback_on_failure applies on the primary path.
* fix(router): restore litellm_params.cooldown_time as a pre-existing fallback
cooldown_time already had litellm_params support on Router.deployment_callback_on_failure
before this PR; the earlier model_info-only restriction (aimed at the leak concern
for the genuinely new allowed_fails/allowed_fails_policy fields) incorrectly dropped
that pre-existing capability too. model_info still takes priority when both are set.
* fix(router): keep the fallback-cooldown trigger in sync with #35104's review fixes
Applies the same two fixes landed on the split-out PR #35104 (which #34416
still duplicates until it's rebased onto the merged base): increment the
deployment's per-minute failure counter before evaluating cooldown, and
require the server-stamped failed_deployment_id instead of trusting a
metadata bucket, since neither "metadata" nor "litellm_metadata" can be told
apart from a caller-supplied one without knowing the call's function_name.
* fix(router): freeze the model_info fallback mapping to satisfy the type-discipline gate
* fix(router): defer f-string interpolation in fallback-cooldown debug logs
* fix(router): annotate cooldown-path locals with Final to satisfy the LIT010 budget
* fix(router): suppress reportPrivateUsage for cross-module cooldown helpers
* fix(router): don't cool down deployments for request-scoped 404s on generic API fallbacks
* fix(router): stamp the dynamic client-side-credential deployment id, not the shared static one
* fix(router): keep up with upstream typing modernization and Final-annotation ratchet
* fix(router): don't cool down deployments for a caller-supplied x-litellm-timeout
* fix(router): stamp dynamic client-side-credential id in completion fallback paths too
The generic-API-call helper already stamped the effective (dynamic-if-client-side-credential)
deployment id on exceptions, but the regular _completion/_acompletion exception handlers still
stamped the static shared deployment's id. A tenant using invalid forwarded credentials could
generate repeated failures attributed to, and eventually cooling down, the shared deployment
other tenants rely on. Extracted the stamping logic into one shared helper used by all three
call sites (generic API, sync completion, async completion) so the fix and future changes to it
stay in one place.
* fix(proxy): recognize body-supplied timeout/request_timeout/stream_timeout as caller-controlled
client_side_timeout was only set when the caller used the x-litellm-timeout header, but
Router._get_timeout also resolves the effective timeout from kwargs["timeout"],
kwargs["request_timeout"], and kwargs["stream_timeout"], all settable directly in the
request body (and x-litellm-stream-timeout wasn't marked either). A caller could set any
of those to a near-zero value, force a 408 on every deployment in a fallback chain, and
cool down deployments other tenants rely on without the guard in
_trigger_cooldown_for_failed_deployment recognizing it as caller-controlled. Also strip
any client-forged client_side_timeout from the request body so the marker is always
server-computed.
---------
Co-authored-by: Deepanshu <deepanshu.lulla@alpha-sense.com>
PR #35956 added the max_input_chars passthrough to the AutoRouter
constructor but left this mock assertion in tests/router_unit_tests
unchanged, so test_init_auto_router_deployment_success has been failing
on litellm_internal_staging ever since.
The passthrough itself is intentional and its behaviour is already
covered by TestAutoRouterMaxInputCharsWiring in tests/test_litellm, so
only the stale expected kwargs need updating. Assert the shared constant
rather than the literal 2000 so tuning the default does not break this
test again.
