Skip to content

Use Claude Code, Codex, VSCode, Pi, and OpenCode (and 6 other harnesses) for free (1.3B+ free tokens) from your terminal, app, IDE, or phone, and now from the browser with native browser sessions (multi-harness + multi-model) like OpenClaw (voice supported + ToS friendly)

Python

patch: Preserve provider protection across settings changes (#1923)

Ali Khokhar37c77262115a
diff --git a/e2e/conftest.py b/e2e/conftest.py
index c9eaa51e..1aed39c5 100644
--- a/e2e/conftest.py
+++ b/e2e/conftest.py
@@ -56,9 +56,6 @@ class _ModelListingProvider(BaseProvider):
ProviderConfig(
api_key="browser-test",
base_url="https://provider.invalid/v1",
- rate_limit=1_000,
- rate_window=1,
- max_concurrency=100,
http_read_timeout=1.0,
http_write_timeout=1.0,
http_connect_timeout=1.0,
@@ -252,15 +249,20 @@ def admin_base_url(
),
}
- async def fixture_provider(provider_id: str, _settings: Settings) -> BaseProvider:
+ async def fixture_provider(
+ provider_id: str, _settings: Settings, _admission_registry
+ ) -> BaseProvider:
if provider_id not in providers:
raise AssertionError(f"Missing browser fixture provider: {provider_id}")
return providers[provider_id]
manager = ProviderRuntimeManager(
get_settings(),
- runtime_factory=lambda snapshot: ProviderRuntime(
- snapshot, dict(providers), provider_constructor=fixture_provider
+ runtime_factory=lambda snapshot, admission_registry: ProviderRuntime(
+ snapshot,
+ admission_registry,
+ dict(providers),
+ provider_constructor=fixture_provider,
),
)
runtime = ApplicationRuntime(
diff --git a/src/free_claude_code/config/settings.py b/src/free_claude_code/config/settings.py
index 6ce7cbfe..00a00a5c 100644
--- a/src/free_claude_code/config/settings.py
+++ b/src/free_claude_code/config/settings.py
@@ -584,12 +584,14 @@ class Settings(BaseModel):
default=None, validation_alias="OLLAMA_CLOUD_PROXY"
)
# ==================== Provider Rate Limiting ====================
- provider_rate_limit: int = Field(default=1, validation_alias="PROVIDER_RATE_LIMIT")
+ provider_rate_limit: int = Field(
+ default=1, gt=0, validation_alias="PROVIDER_RATE_LIMIT"
+ )
provider_rate_window: int = Field(
- default=2, validation_alias="PROVIDER_RATE_WINDOW"
+ default=2, gt=0, validation_alias="PROVIDER_RATE_WINDOW"
)
provider_max_concurrency: int = Field(
- default=2, validation_alias="PROVIDER_MAX_CONCURRENCY"
+ default=2, gt=0, validation_alias="PROVIDER_MAX_CONCURRENCY"
)
provider_progress_timeout: float = Field(
default=600.0,
diff --git a/src/free_claude_code/providers/admission.py b/src/free_claude_code/providers/admission.py
index 3bd9490a..51f3373f 100644
--- a/src/free_claude_code/providers/admission.py
+++ b/src/free_claude_code/providers/admission.py
@@ -5,6 +5,7 @@ import math
import random
import time
import uuid
+from collections import deque
from collections.abc import Awaitable, Callable
from dataclasses import dataclass, field
from datetime import UTC, datetime
@@ -14,8 +15,8 @@ from typing import TypeVar
from loguru import logger
-from free_claude_code.core.rate_limit import StrictSlidingWindowLimiter
from free_claude_code.core.trace import trace_event
+from free_claude_code.providers.admission_policy import ProviderAdmissionLimits
from free_claude_code.providers.failure_policy import (
ProviderFailureOverride,
ProviderRecoveryExhausted,
@@ -416,12 +417,7 @@ class ProviderAdmissionController:
max_delay: float = DEFAULT_UPSTREAM_MAX_DELAY,
jitter: float = DEFAULT_UPSTREAM_JITTER,
) -> None:
- if rate_limit <= 0:
- raise ValueError("rate_limit must be > 0")
- if rate_window <= 0:
- raise ValueError("rate_window must be > 0")
- if max_concurrency <= 0:
- raise ValueError("max_concurrency must be > 0")
+ limits = ProviderAdmissionLimits(rate_limit, rate_window, max_concurrency)
if max_attempts <= 0:
raise ValueError("max_attempts must be > 0")
if base_delay < 0:
@@ -436,10 +432,11 @@ class ProviderAdmissionController:
self._base_delay = base_delay
self._max_delay = max_delay
self._jitter = jitter
- self._proactive_limiter = StrictSlidingWindowLimiter(
- rate_limit, float(rate_window)
- )
- self._concurrency_sem = asyncio.Semaphore(max_concurrency)
+ self._limits = limits
+ self._admission_times: deque[float] = deque()
+ self._last_discarded_at: float | None = None
+ self._active_attempts = 0
+ self._capacity_changed = asyncio.Event()
self._condition = asyncio.Condition()
self._episode: _RecoveryEpisode | None = None
self._next_generation = 1
@@ -485,19 +482,11 @@ class ProviderAdmissionController:
slot_acquired = False
claim: _AttemptClaim | None = None
try:
- admitted = await self._proactive_limiter.acquire_if(
- lambda permit=permit: self._permit_is_current(execution, permit)
- )
+ admitted = await self._acquire_capacity(execution, permit)
if not admitted:
await self._abandon_probe_permit(execution, permit)
continue
- await self._concurrency_sem.acquire()
slot_acquired = True
- if not self._permit_is_current(execution, permit):
- self._concurrency_sem.release()
- slot_acquired = False
- await self._abandon_probe_permit(execution, permit)
- continue
claim = execution._claim_attempt(operation_kind)
attempt = ProviderAttempt(self, execution, permit, claim)
trace_event(
@@ -518,7 +507,7 @@ class ProviderAdmissionController:
if claim is not None:
execution._close_attempt(claim)
if slot_acquired:
- self._concurrency_sem.release()
+ self._release_concurrency()
await self._abandon_probe_permit(execution, permit)
raise
@@ -639,7 +628,7 @@ class ProviderAdmissionController:
if episode is None:
return
self._episode = None
- self._condition.notify_all()
+ self._notify_recovery_waiters()
trace_event(
stage="provider",
event="provider.recovery.closed",
@@ -664,7 +653,7 @@ class ProviderAdmissionController:
return
episode.probe_active = False
episode.ready_at = time.monotonic()
- self._condition.notify_all()
+ self._notify_recovery_waiters()
async def _attempt_rejected(
self,
@@ -679,7 +668,7 @@ class ProviderAdmissionController:
if episode is None:
return
self._episode = None
- self._condition.notify_all()
+ self._notify_recovery_waiters()
trace_event(
stage="provider",
event="provider.recovery.closed",
@@ -705,7 +694,7 @@ class ProviderAdmissionController:
episode.leader = None
episode.probe_active = False
episode.ready_at = time.monotonic()
- self._condition.notify_all()
+ self._notify_recovery_waiters()
async def _attempt_failed(
self,
@@ -740,7 +729,7 @@ class ProviderAdmissionController:
execution._fail_recovery(episode.last_error)
else:
episode.waiters.add(execution)
- self._condition.notify_all()
+ self._notify_recovery_waiters()
elif episode is None or matching_probe is not None:
terminal_delay = self._retry_delay(
error,
@@ -761,7 +750,7 @@ class ProviderAdmissionController:
waiter._fail_recovery(error)
episode.waiters.clear()
exhausted_episode = True
- self._condition.notify_all()
+ self._notify_recovery_waiters()
label = self._failure_label(status, error)
if became_leader:
@@ -871,7 +860,7 @@ class ProviderAdmissionController:
episode.leader = None
episode.probe_active = False
episode.ready_at = time.monotonic()
- self._condition.notify_all()
+ self._notify_recovery_waiters()
async def _abandon_waiting_leader(
self,
@@ -888,10 +877,64 @@ class ProviderAdmissionController:
):
return
episode.leader = None
- self._condition.notify_all()
+ self._notify_recovery_waiters()
+
+ def reconfigure(self, limits: ProviderAdmissionLimits) -> None:
+ """Update capacity without resetting occupancy, history, or recovery."""
+ if limits != self._limits:
+ self._limits = limits
+ self._wake_capacity()
+
+ def _wake_capacity(self) -> None:
+ changed = self._capacity_changed
+ self._capacity_changed = asyncio.Event()
+ changed.set()
+
+ def _notify_recovery_waiters(self) -> None:
+ self._condition.notify_all()
+ self._wake_capacity()
+
+ async def _acquire_capacity(
+ self, execution: ProviderExecution, permit: _GatePermit
+ ) -> bool:
+ while self._permit_is_current(execution, permit):
+ changed = self._capacity_changed
+ now = time.monotonic()
+ limits = self._limits
+ cutoff = now - limits.rate_window
+ while self._admission_times and self._admission_times[0] <= cutoff:
+ self._last_discarded_at = self._admission_times.popleft()
+
+ ready_at = now
+ if self._last_discarded_at is not None:
+ ready_at = max(ready_at, self._last_discarded_at + limits.rate_window)
+ if len(self._admission_times) >= limits.rate_limit:
+ ready_at = max(
+ ready_at,
+ self._admission_times[-limits.rate_limit] + limits.rate_window,
+ )
+ if self._active_attempts < limits.max_concurrency and ready_at <= now:
+ self._admission_times.append(now)
+ self._active_attempts += 1
+ return True
+
+ # Occupancy changes explicitly wake us. Avoid an expired rate timer
+ # spinning while all concurrency slots remain occupied.
+ delay = (
+ max(0.0, ready_at - now)
+ if self._active_attempts < limits.max_concurrency
+ else None
+ )
+ try:
+ async with asyncio.timeout(delay):
+ await changed.wait()
+ except TimeoutError:
+ pass
+ return False
def _release_concurrency(self) -> None:
- self._concurrency_sem.release()
+ self._active_attempts -= 1
+ self._wake_capacity()
def _retry_delay(self, error: Exception, attempt: int) -> float:
exponent = max(0, attempt - 1)
diff --git a/src/free_claude_code/providers/admission_policy.py b/src/free_claude_code/providers/admission_policy.py
new file mode 100644
index 00000000..3b423d8f
--- /dev/null
+++ b/src/free_claude_code/providers/admission_policy.py
@@ -0,0 +1,31 @@
+"""Validated capacity policy, independent of clients and controller lifetime."""
+
+import math
+from dataclasses import dataclass
+from typing import TYPE_CHECKING
+
+if TYPE_CHECKING:
+ from free_claude_code.config.settings import Settings
+
+
+@dataclass(frozen=True, slots=True)
+class ProviderAdmissionLimits:
+ rate_limit: int
+ rate_window: float
+ max_concurrency: int
+
+ def __post_init__(self) -> None:
+ for name in ("rate_limit", "max_concurrency"):
+ value = getattr(self, name)
+ if not isinstance(value, int) or isinstance(value, bool) or value <= 0:
+ raise ValueError(f"{name} must be a positive integer")
+ if not math.isfinite(self.rate_window) or self.rate_window <= 0:
+ raise ValueError("rate_window must be finite and > 0")
+
+ @classmethod
+ def from_settings(cls, settings: Settings) -> ProviderAdmissionLimits:
+ return cls(
+ settings.provider_rate_limit,
+ settings.provider_rate_window,
+ settings.provider_max_concurrency,
+ )
diff --git a/src/free_claude_code/providers/admission_registry.py b/src/free_claude_code/providers/admission_registry.py
new file mode 100644
index 00000000..772c11bf
--- /dev/null
+++ b/src/free_claude_code/providers/admission_registry.py
@@ -0,0 +1,50 @@
+"""Loop-owned admission policy and controller lifetimes, independent of clients."""
+
+from collections.abc import Iterable
+from typing import TYPE_CHECKING
+
+from free_claude_code.providers.admission_policy import ProviderAdmissionLimits
+
+if TYPE_CHECKING:
+ from free_claude_code.providers.admission import ProviderAdmissionController
+
+
+class ProviderAdmissionRegistry:
+ """One protection budget per configured provider for one runtime manager."""
+
+ def __init__(self, limits: ProviderAdmissionLimits) -> None:
+ self._limits = limits
+ self._controllers: dict[str, ProviderAdmissionController] = {}
+ self._closed = False
+
+ def get(self, provider_id: str) -> ProviderAdmissionController:
+ if self._closed:
+ raise RuntimeError("Provider admission registry is closed")
+ if provider_id not in self._controllers:
+ # Provider preparation loads SDKs in a worker before reaching here.
+ from free_claude_code.providers.admission import ProviderAdmissionController
+
+ self._controllers[provider_id] = ProviderAdmissionController(
+ provider_name=provider_id,
+ rate_limit=self._limits.rate_limit,
+ rate_window=self._limits.rate_window,
+ max_concurrency=self._limits.max_concurrency,
+ )
+ return self._controllers[provider_id]
+
+ def reconfigure(self, limits: ProviderAdmissionLimits) -> None:
+ """Publish validated limits without yielding to waiting requests."""
+ self._limits = limits
+ for controller in self._controllers.values():
+ controller.reconfigure(limits)
+
+ def retain_custom(self, provider_ids: Iterable[str]) -> None:
+ retained = set(provider_ids)
+ for provider_id in tuple(self._controllers):
+ if provider_id.startswith("custom_") and provider_id not in retained:
+ del self._controllers[provider_id]
+
+ def close(self) -> None:
+ """Forget state only after client generations have drained and closed."""
+ self._closed = True
+ self._controllers.clear()
diff --git a/src/free_claude_code/providers/base.py b/src/free_claude_code/providers/base.py
index ecf11ab8..a8ab4405 100644
--- a/src/free_claude_code/providers/base.py
+++ b/src/free_claude_code/providers/base.py
@@ -20,9 +20,6 @@ class ProviderConfig:
api_key: str | None
base_url: str
- rate_limit: int
- rate_window: int
- max_concurrency: int
http_read_timeout: float
http_write_timeout: float
http_connect_timeout: float
diff --git a/src/free_claude_code/providers/runtime/config.py b/src/free_claude_code/providers/runtime/config.py
index a36e9b92..0c6a3dd4 100644
--- a/src/free_claude_code/providers/runtime/config.py
+++ b/src/free_claude_code/providers/runtime/config.py
@@ -80,9 +80,6 @@ def build_provider_config(
return ProviderConfig(
api_key=credential,
base_url=resolved_base_url,
- rate_limit=settings.provider_rate_limit,
- rate_window=settings.provider_rate_window,
- max_concurrency=settings.provider_max_concurrency,
http_read_timeout=settings.http_read_timeout,
http_write_timeout=settings.http_write_timeout,
http_connect_timeout=settings.http_connect_timeout,
@@ -99,9 +96,6 @@ def build_custom_provider_config(
api_key=definition.api_key.get_secret_value() if definition.api_key else None,
base_url=definition.base_url,
proxy=None,
- rate_limit=settings.provider_rate_limit,
- rate_window=settings.provider_rate_window,
- max_concurrency=settings.provider_max_concurrency,
http_read_timeout=settings.http_read_timeout,
http_write_timeout=settings.http_write_timeout,
http_connect_timeout=settings.http_connect_timeout,
diff --git a/src/free_claude_code/providers/runtime/factory.py b/src/free_claude_code/providers/runtime/factory.py
index cd08d7b9..95334512 100644
--- a/src/free_claude_code/providers/runtime/factory.py
+++ b/src/free_claude_code/providers/runtime/factory.py
@@ -11,6 +11,7 @@ from free_claude_code.config.custom_providers import CustomProviderDefinition
from free_claude_code.config.provider_catalog import PROVIDER_CATALOG
from free_claude_code.config.settings import Settings
from free_claude_code.providers.admission import ProviderAdmissionController
+from free_claude_code.providers.admission_registry import ProviderAdmissionRegistry
from free_claude_code.providers.base import BaseProvider, ProviderConfig
from free_claude_code.providers.openai_chat import (
OPENAI_CHAT_PROFILES,
@@ -251,7 +252,7 @@ def prepare_provider(
provider_id: str,
provider_loaders: Mapping[str, Callable[[], ProviderFactory]],
custom_definition: CustomProviderDefinition | None = None,
-) -> Callable[[Settings], BaseProvider]:
+) -> Callable[[Settings, ProviderAdmissionRegistry], BaseProvider]:
"""Load implementation modules in a worker; return a loop-owned constructor."""
# The SDK lazily imports these on first client resource access. Keep that
@@ -262,14 +263,11 @@ def prepare_provider(
from .config import build_custom_provider_config
- def construct_custom(settings: Settings) -> BaseProvider:
+ def construct_custom(
+ settings: Settings, admission_registry: ProviderAdmissionRegistry
+ ) -> BaseProvider:
config = build_custom_provider_config(custom_definition, settings)
- admission = ProviderAdmissionController(
- provider_name=provider_id,
- rate_limit=config.rate_limit,
- rate_window=config.rate_window,
- max_concurrency=config.max_concurrency,
- )
+ admission = admission_registry.get(provider_id)
return CustomProvider(
config, definition=custom_definition, admission=admission
)
@@ -287,14 +285,11 @@ def prepare_provider(
)
factory = loader() if loader is not None else None
- def construct(settings: Settings) -> BaseProvider:
+ def construct(
+ settings: Settings, admission_registry: ProviderAdmissionRegistry
+ ) -> BaseProvider:
config = build_provider_config(descriptor, settings)
- admission = ProviderAdmissionController(
- provider_name=provider_id,
- rate_limit=config.rate_limit,
- rate_window=config.rate_window,
- max_concurrency=config.max_concurrency,
- )
+ admission = admission_registry.get(provider_id)
if factory is not None:
return factory(config, settings, admission)
return create_openai_chat_provider(provider_id, config, admission)
diff --git a/src/free_claude_code/providers/runtime/runtime.py b/src/free_claude_code/providers/runtime/runtime.py
index 0a2ebd09..a615af40 100644
--- a/src/free_claude_code/providers/runtime/runtime.py
+++ b/src/free_claude_code/providers/runtime/runtime.py
@@ -8,18 +8,21 @@ from typing import TYPE_CHECKING
from free_claude_code.application.errors import ApplicationUnavailableError
from free_claude_code.config.settings import Settings
from free_claude_code.core.async_tasks import run_sync_owned
+from free_claude_code.providers.admission_registry import ProviderAdmissionRegistry
from free_claude_code.providers.base import BaseProvider
if TYPE_CHECKING:
from .factory import ProviderFactory
-type ProviderConstructor = Callable[[str, Settings], Awaitable[BaseProvider]]
+type ProviderConstructor = Callable[
+ [str, Settings, ProviderAdmissionRegistry], Awaitable[BaseProvider]
+]
type ProviderLoader = Callable[[], ProviderFactory]
def _load_constructor(
provider_id: str, provider_loaders: Mapping[str, ProviderLoader], settings: Settings
-) -> Callable[[Settings], BaseProvider]:
+) -> Callable[[Settings, ProviderAdmissionRegistry], BaseProvider]:
from .factory import prepare_provider
return prepare_provider(
@@ -34,13 +37,14 @@ def _load_constructor(
async def create_provider(
provider_id: str,
settings: Settings,
+ admission_registry: ProviderAdmissionRegistry,
*,
provider_loaders: Mapping[str, ProviderLoader] | None = None,
) -> BaseProvider:
constructor = await run_sync_owned(
partial(_load_constructor, provider_id, provider_loaders or {}, settings)
)
- return constructor(settings)
+ return constructor(settings, admission_registry)
class ProviderRuntime:
@@ -49,11 +53,13 @@ class ProviderRuntime:
def __init__(
self,
settings: Settings,
+ admission_registry: ProviderAdmissionRegistry,
providers: MutableMapping[str, BaseProvider] | None = None,
*,
provider_constructor: ProviderConstructor = create_provider,
) -> None:
self.settings = settings
+ self._admission_registry = admission_registry
self._providers = providers if providers is not None else {}
self._provider_constructor = provider_constructor
self._creations: dict[str, asyncio.Task[BaseProvider]] = {}
@@ -85,7 +91,9 @@ class ProviderRuntime:
task.exception() # Observe failures even when every caller stopped waiting.
async def _construct(self, provider_id: str) -> BaseProvider:
- provider = await self._provider_constructor(provider_id, self.settings)
+ provider = await self._provider_constructor(
+ provider_id, self.settings, self._admission_registry
+ )
self._providers[provider_id] = provider
return provider
diff --git a/src/free_claude_code/runtime/provider_manager.py b/src/free_claude_code/runtime/provider_manager.py
index 7b124d6e..e04bc9e8 100644
--- a/src/free_claude_code/runtime/provider_manager.py
+++ b/src/free_claude_code/runtime/provider_manager.py
@@ -20,6 +20,8 @@ from free_claude_code.core.async_tasks import run_sync_owned
from free_claude_code.core.json_types import JsonObject
from free_claude_code.core.token_estimation import initialize_token_estimation
from free_claude_code.core.trace import trace_event
+from free_claude_code.providers.admission_policy import ProviderAdmissionLimits
+from free_claude_code.providers.admission_registry import ProviderAdmissionRegistry
from free_claude_code.providers.base import BaseProvider
from free_claude_code.providers.model_listing import model_infos_from_ids
from free_claude_code.providers.runtime.discovery import (
@@ -31,7 +33,9 @@ from free_claude_code.providers.runtime.discovery import (
from free_claude_code.providers.runtime.model_cache import ProviderModelCache
from free_claude_code.providers.runtime.runtime import ProviderRuntime
-ProviderRuntimeFactory = Callable[[Settings], ProviderRuntime]
+ProviderRuntimeFactory = Callable[
+ [Settings, ProviderAdmissionRegistry], ProviderRuntime
+]
ConnectedProviderIds = Callable[[], tuple[str, ...]]
CommitConfig = Callable[[], Awaitable[None]]
@@ -157,6 +161,9 @@ class ProviderRuntimeManager:
model_catalog_publisher: ModelCatalogPublisher | None = None,
) -> None:
self._runtime_factory = runtime_factory
+ self._admission_registry = ProviderAdmissionRegistry(
+ ProviderAdmissionLimits.from_settings(settings)
+ )
self._connected_provider_ids = connected_provider_ids
self._model_catalog_publisher = model_catalog_publisher
self._catalog_changed: Callable[[], None] | None = None
@@ -174,7 +181,7 @@ class ProviderRuntimeManager:
self._current = _ProviderGeneration(
generation_id=1,
settings=settings,
- runtime=runtime_factory(settings),
+ runtime=runtime_factory(settings, self._admission_registry),
cache=ProviderModelCache(
model_cache_provider_ids_for_settings(
settings, connected_provider_ids()
@@ -527,10 +534,13 @@ class ProviderRuntimeManager:
async with self._replace_lock:
self._ensure_open()
await self._retry_unpublished_cleanup()
+ limits = ProviderAdmissionLimits.from_settings(settings)
candidate_id = self._next_generation_id
candidate_runtime: ProviderRuntime | None = None
try:
- candidate_runtime = self._runtime_factory(settings)
+ candidate_runtime = self._runtime_factory(
+ settings, self._admission_registry
+ )
await commit()
except BaseException as exc:
trace_event(
@@ -565,6 +575,7 @@ class ProviderRuntimeManager:
runtime=candidate_runtime,
cache=cache,
)
+ self._admission_registry.reconfigure(limits)
self._current = candidate
self._catalog_revision += 1
previous.retired = True
@@ -621,8 +632,21 @@ class ProviderRuntimeManager:
if not all(generation_results) or not unpublished_closed:
raise RuntimeError("One or more provider runtimes failed to close.")
self._current.cache.clear()
+ self._admission_registry.close()
self._closed = True
+ def _prune_admission(self) -> None:
+ runtimes = (
+ self._current.runtime,
+ *(generation.runtime for generation in self._retired.values()),
+ *self._unpublished,
+ )
+ self._admission_registry.retain_custom(
+ definition.provider_id
+ for runtime in runtimes
+ for definition in runtime.settings.custom_providers
+ )
+
async def _release(self, generation: _ProviderGeneration) -> None:
if generation.active_leases <= 0:
return
@@ -645,6 +669,7 @@ class ProviderRuntimeManager:
)
return False
self._unpublished.discard(runtime)
+ self._prune_admission()
return True
async def _retry_unpublished_cleanup(self) -> bool:
@@ -696,6 +721,7 @@ class ProviderRuntimeManager:
generation.closed = True
self._retired.pop(generation.generation_id, None)
+ self._prune_admission()
trace_event(
stage="runtime",
event="provider_generation.closed",
diff --git a/tests/api/support.py b/tests/api/support.py
index ff7494da..c5543727 100644
--- a/tests/api/support.py
+++ b/tests/api/support.py
@@ -67,8 +67,9 @@ def create_test_app(
else:
manager = ApiTestRuntime(
settings,
- runtime_factory=lambda snapshot: ProviderRuntime(
+ runtime_factory=lambda snapshot, admission_registry: ProviderRuntime(
snapshot,
+ admission_registry,
dict(providers),
),
connected_provider_ids=connected_provider_ids,
diff --git a/tests/api/test_openai_codex_compatibility.py b/tests/api/test_openai_codex_compatibility.py
index d1e15240..2d48f0ad 100644
--- a/tests/api/test_openai_codex_compatibility.py
+++ b/tests/api/test_openai_codex_compatibility.py
@@ -71,9 +71,6 @@ def _openai_provider(
make_provider_config(
api_key="",
base_url="https://chatgpt.com/backend-api/codex",
- rate_limit=100,
- rate_window=1,
- max_concurrency=2,
),
auth=_FakeOpenAIAuth(),
admission=ProviderAdmissionController(
diff --git a/tests/cli/test_config_restart_lifecycle.py b/tests/cli/test_config_restart_lifecycle.py
index 404a58a3..d01ed8ab 100644
--- a/tests/cli/test_config_restart_lifecycle.py
+++ b/tests/cli/test_config_restart_lifecycle.py
@@ -35,7 +35,10 @@ def test_supervised_http_apply_finishes_and_reconnects(monkeypatch, stop_during_
def build(settings, restart_callback):
manager = ProviderRuntimeManager(
- settings, runtime_factory=lambda snapshot: ProviderRuntime(snapshot, {})
+ settings,
+ runtime_factory=lambda snapshot, admission_registry: ProviderRuntime(
+ snapshot, admission_registry, {}
+ ),
)
monkeypatch.setattr(manager, "start_model_list_refresh", lambda: None)
monkeypatch.setattr(manager, "_start_pass", lambda *args, **kwargs: None)
diff --git a/tests/conftest.py b/tests/conftest.py
index 3da3ad97..e2799bf4 100644
--- a/tests/conftest.py
+++ b/tests/conftest.py
@@ -85,9 +85,6 @@ def provider_config():
return make_provider_config(
api_key="test_key",
base_url="https://test.api.nvidia.com/v1",
- rate_limit=10,
- rate_window=60,
- max_concurrency=5,
http_read_timeout=300.0,
http_write_timeout=10.0,
http_connect_timeout=10.0,
@@ -123,9 +120,6 @@ def lmstudio_provider(provider_config):
lmstudio_config = make_provider_config(
api_key="lm-studio",
base_url="http://localhost:1234/v1",
- rate_limit=provider_config.rate_limit,
- rate_window=provider_config.rate_window,
- max_concurrency=provider_config.max_concurrency,
http_read_timeout=provider_config.http_read_timeout,
http_write_timeout=provider_config.http_write_timeout,
http_connect_timeout=provider_config.http_connect_timeout,
@@ -143,9 +137,6 @@ def llamacpp_provider(provider_config):
llamacpp_config = make_provider_config(
api_key="llamacpp",
base_url="http://localhost:8080/v1",
- rate_limit=10,
- rate_window=60,
- max_concurrency=5,
http_read_timeout=300.0,
http_write_timeout=10.0,
http_connect_timeout=10.0,
diff --git a/tests/contracts/test_startup_import_boundaries.py b/tests/contracts/test_startup_import_boundaries.py
index 0bb62026..f76fd192 100644
--- a/tests/contracts/test_startup_import_boundaries.py
+++ b/tests/contracts/test_startup_import_boundaries.py
@@ -41,6 +41,8 @@ from unittest.mock import MagicMock
from free_claude_code.config.settings import Settings
from free_claude_code.providers.base import BaseProvider
from free_claude_code.providers.runtime.runtime import create_provider
+from free_claude_code.providers.admission_policy import ProviderAdmissionLimits
+from free_claude_code.providers.admission_registry import ProviderAdmissionRegistry
async def main():
loop = asyncio.get_running_loop()
@@ -54,7 +56,7 @@ async def main():
assert threading.get_ident() == thread
return provider
return construct
- assert await create_provider("groq", Settings(groq_api_key="test"), provider_loaders={"groq": load}) is provider
+ assert await create_provider("groq", Settings(groq_api_key="test"), ProviderAdmissionRegistry(ProviderAdmissionLimits(1, 2, 2)), provider_loaders={"groq": load}) is provider
await provider.cleanup()
asyncio.run(main())
diff --git a/tests/providers/support.py b/tests/providers/support.py
index 63cf6fde..4da65414 100644
--- a/tests/providers/support.py
+++ b/tests/providers/support.py
@@ -52,9 +52,6 @@ class SDKStreamDouble[EventT](AsyncIterator[EventT]):
def make_provider_config(
api_key: str | None,
base_url: str,
- rate_limit: int = 1_000_000,
- rate_window: int = 1,
- max_concurrency: int = 1_000,
http_read_timeout: float = 120.0,
http_write_timeout: float = 10.0,
http_connect_timeout: float = 10.0,
@@ -67,9 +64,6 @@ def make_provider_config(
return ProviderConfig(
api_key=api_key,
base_url=base_url,
- rate_limit=rate_limit,
- rate_window=rate_window,
- max_concurrency=max_concurrency,
http_read_timeout=http_read_timeout,
http_write_timeout=http_write_timeout,
http_connect_timeout=http_connect_timeout,
diff --git a/tests/providers/test_admission_reconfiguration.py b/tests/providers/test_admission_reconfiguration.py
new file mode 100644
index 00000000..1d0b025e
--- /dev/null
+++ b/tests/providers/test_admission_reconfiguration.py
@@ -0,0 +1,146 @@
+"""Live policy changes preserve capacity and wake the correct waiting calls."""
+
+import asyncio
+from types import SimpleNamespace
+
+import pytest
+from pydantic import ValidationError
+
+from free_claude_code.config.settings import Settings
+from free_claude_code.providers import admission as admission_module
+from free_claude_code.providers.admission import (
+ ProviderAdmissionController,
+ ProviderOperationKind,
+)
+from free_claude_code.providers.admission_policy import ProviderAdmissionLimits
+
+
+async def admit(controller):
+ return await controller.start_execution().open_attempt(
+ ProviderOperationKind.GENERATION
+ )
+
+
+async def completed(controller):
+ attempt = await admit(controller)
+ await attempt.accept()
+ await attempt.aclose()
+
+
+async def test_limit_changes_preserve_occupied_slots_and_wake_waiters():
+ controller = ProviderAdmissionController(
+ provider_name="test", rate_limit=1000, max_concurrency=1
+ )
+ first = await admit(controller)
+ pending = asyncio.create_task(admit(controller))
+ await asyncio.sleep(0)
+ assert not pending.done()
+ controller.reconfigure(ProviderAdmissionLimits(1000, 60, 2))
+ second = await asyncio.wait_for(pending, 1)
+ controller.reconfigure(ProviderAdmissionLimits(1000, 60, 1))
+ third = asyncio.create_task(admit(controller))
+ await first.aclose()
+ await asyncio.sleep(0)
+ assert not third.done()
+ await second.aclose()
+ await (await asyncio.wait_for(third, 1)).aclose()
+
+
+async def test_waiting_for_concurrency_does_not_spend_rate_capacity(monkeypatch):
+ clock = SimpleNamespace(now=0.0)
+ monkeypatch.setattr(
+ admission_module, "time", SimpleNamespace(monotonic=lambda: clock.now)
+ )
+ controller = ProviderAdmissionController(
+ provider_name="test", rate_limit=1, rate_window=10, max_concurrency=1
+ )
+ first = await admit(controller)
+ clock.now = 11
+ pending = asyncio.create_task(admit(controller))
+ await asyncio.sleep(0)
+ pending.cancel()
+ await asyncio.gather(pending, return_exceptions=True)
+ await first.aclose()
+ await asyncio.wait_for(completed(controller), 1)
+
+
+async def test_rate_count_increase_wakes_waiter_without_clearing_history():
+ controller = ProviderAdmissionController(
+ provider_name="test", rate_limit=1, rate_window=60
+ )
+ await completed(controller)
+ pending = asyncio.create_task(admit(controller))
+ await asyncio.sleep(0)
+ assert not pending.done()
+ controller.reconfigure(ProviderAdmissionLimits(2, 60, 5))
+ await (await asyncio.wait_for(pending, 1)).aclose()
+ blocked = asyncio.create_task(admit(controller))
+ await asyncio.sleep(0)
+ assert not blocked.done()
+ blocked.cancel()
+ await asyncio.gather(blocked, return_exceptions=True)
+
+
+async def test_longer_window_waits_only_until_discarded_history_expires(monkeypatch):
+ clock = SimpleNamespace(now=0.0)
+ monkeypatch.setattr(
+ admission_module, "time", SimpleNamespace(monotonic=lambda: clock.now)
+ )
+ controller = ProviderAdmissionController(
+ provider_name="test", rate_limit=1, rate_window=2
+ )
+ await completed(controller)
+ clock.now = 3
+ await completed(controller) # Drops the timestamp at zero.
+ controller.reconfigure(ProviderAdmissionLimits(100, 10, 5))
+ pending = asyncio.create_task(admit(controller))
+ await asyncio.sleep(0)
+ assert not pending.done() # Capacity is ample, but older history is incomplete.
+ clock.now = 9
+ controller.reconfigure(ProviderAdmissionLimits(101, 10, 5))
+ await asyncio.sleep(0)
+ assert not pending.done()
+ clock.now = 10
+ controller.reconfigure(ProviderAdmissionLimits(102, 10, 5))
+ await (await asyncio.wait_for(pending, 1)).aclose()
+
+
+async def test_shorter_window_wakes_waiter_and_longer_window_uses_retained_history(
+ monkeypatch,
+):
+ clock = SimpleNamespace(now=0.0)
+ monkeypatch.setattr(
+ admission_module, "time", SimpleNamespace(monotonic=lambda: clock.now)
+ )
+ controller = ProviderAdmissionController(
+ provider_name="test", rate_limit=1, rate_window=60
+ )
+ await completed(controller)
+ controller.reconfigure(ProviderAdmissionLimits(2, 120, 5))
+ await asyncio.wait_for(completed(controller), 1) # No artificial expansion pause.
+ pending = asyncio.create_task(admit(controller))
+ await asyncio.sleep(0)
+ assert not pending.done()
+ clock.now = 2
+ controller.reconfigure(ProviderAdmissionLimits(1, 2, 5))
+ await (await asyncio.wait_for(pending, 1)).aclose()
+
+
+ "field", ["provider_rate_limit", "provider_rate_window", "provider_max_concurrency"]
+)
[email protected]("value", [0, -1])
+def test_settings_reject_invalid_admission_limits(field, value):
+ with pytest.raises(ValidationError):
+ Settings.model_validate({field: value})
+
+
[email protected]("window", [float("inf"), float("nan"), 0, -1])
+def test_internal_policy_rejects_invalid_window(window):
+ with pytest.raises(ValueError, match="rate_window"):
+ ProviderAdmissionLimits(1, window, 1)
diff --git a/tests/providers/test_agnes.py b/tests/providers/test_agnes.py
index 77dc384b..f4640bd4 100644
--- a/tests/providers/test_agnes.py
+++ b/tests/providers/test_agnes.py
@@ -31,8 +31,6 @@ def agnes_provider() -> OpenAIChatProvider:
make_provider_config(
api_key="test-agnes-key",
base_url=AGNES_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(provider_name="agnes"),
)
@@ -211,8 +209,6 @@ async def test_model_catalog_uses_documented_endpoint_and_auth() -> None:
make_provider_config(
api_key="wire-agnes-key",
base_url=AGNES_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(provider_name="agnes"),
)
diff --git a/tests/providers/test_cerebras.py b/tests/providers/test_cerebras.py
index 501633e2..e8f32311 100644
--- a/tests/providers/test_cerebras.py
+++ b/tests/providers/test_cerebras.py
@@ -57,8 +57,6 @@ def cerebras_config():
return make_provider_config(
api_key="test_cerebras_key",
base_url=CEREBRAS_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
)
@@ -127,8 +125,6 @@ def test_replay_is_independent_of_current_turn_reasoning_control():
make_provider_config(
api_key="test_cerebras_key",
base_url=CEREBRAS_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(),
)
diff --git a/tests/providers/test_cheaperinference.py b/tests/providers/test_cheaperinference.py
index bf71d9d4..f2ecc4fa 100644
--- a/tests/providers/test_cheaperinference.py
+++ b/tests/providers/test_cheaperinference.py
@@ -32,8 +32,6 @@ def cheaperinference_provider() -> OpenAIChatProvider:
make_provider_config(
api_key="test-cheaperinference-key",
base_url=CHEAPERINFERENCE_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(provider_name="cheaperinference", max_attempts=1),
)
diff --git a/tests/providers/test_chutes.py b/tests/providers/test_chutes.py
index eb8692e3..2c5d5f48 100644
--- a/tests/providers/test_chutes.py
+++ b/tests/providers/test_chutes.py
@@ -32,8 +32,6 @@ def chutes_provider() -> OpenAIChatProvider:
make_provider_config(
api_key="test-chutes-key",
base_url=CHUTES_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(provider_name="chutes"),
)
diff --git a/tests/providers/test_cline_pass.py b/tests/providers/test_cline_pass.py
index 2c982307..4cad67e1 100644
--- a/tests/providers/test_cline_pass.py
+++ b/tests/providers/test_cline_pass.py
@@ -40,8 +40,6 @@ def cline_pass_provider() -> OpenAIChatProvider:
make_provider_config(
api_key="test-cline-key",
base_url=CLINE_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(provider_name="cline_pass"),
)
@@ -67,8 +65,6 @@ def _provider_with_transport(
make_provider_config(
api_key=api_key,
base_url=CLINE_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(provider_name="cline_pass"),
)
diff --git a/tests/providers/test_cloudflare.py b/tests/providers/test_cloudflare.py
index 7bcff676..4e7c9086 100644
--- a/tests/providers/test_cloudflare.py
+++ b/tests/providers/test_cloudflare.py
@@ -34,8 +34,6 @@ def cloudflare_config() -> ProviderConfig:
return make_provider_config(
api_key="test-cloudflare-token",
base_url=CLOUDFLARE_AI_REST_ROOT,
- rate_limit=10,
- rate_window=60,
)
diff --git a/tests/providers/test_codestral.py b/tests/providers/test_codestral.py
index 18bd740b..bb1085b9 100644
--- a/tests/providers/test_codestral.py
+++ b/tests/providers/test_codestral.py
@@ -25,8 +25,6 @@ def codestral_config():
return make_provider_config(
api_key="test_codestral_key",
base_url=CODESTRAL_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
)
@@ -72,8 +70,6 @@ def test_build_request_body_global_disable_blocks_reasoning_mapping():
make_provider_config(
api_key="test_codestral_key",
base_url=CODESTRAL_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(),
)
diff --git a/tests/providers/test_cohere.py b/tests/providers/test_cohere.py
index 2e5acb90..27b6a56f 100644
--- a/tests/providers/test_cohere.py
+++ b/tests/providers/test_cohere.py
@@ -26,8 +26,6 @@ def cohere_config():
return make_provider_config(
api_key="test_cohere_key",
base_url=COHERE_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
)
@@ -147,8 +145,6 @@ def test_build_request_body_maps_reasoning_off_to_none():
make_provider_config(
api_key="test_cohere_key",
base_url=COHERE_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(),
)
diff --git a/tests/providers/test_deepinfra.py b/tests/providers/test_deepinfra.py
index b2734849..c8da2853 100644
--- a/tests/providers/test_deepinfra.py
+++ b/tests/providers/test_deepinfra.py
@@ -71,8 +71,6 @@ def deepinfra_provider() -> OpenAIChatProvider:
make_provider_config(
api_key="test-deepinfra-key",
base_url=DEEPINFRA_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(provider_name="deepinfra"),
)
@@ -290,8 +288,6 @@ async def test_model_catalog_uses_absolute_public_url() -> None:
make_provider_config(
api_key="wire-deepinfra-key",
base_url=DEEPINFRA_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(provider_name="deepinfra"),
)
diff --git a/tests/providers/test_deepseek.py b/tests/providers/test_deepseek.py
index ae03d119..78730315 100644
--- a/tests/providers/test_deepseek.py
+++ b/tests/providers/test_deepseek.py
@@ -40,8 +40,6 @@ def deepseek_config():
return make_provider_config(
api_key="test_deepseek_key",
base_url=DEEPSEEK_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
)
@@ -368,8 +366,6 @@ def test_build_request_body_encodes_reasoning_off():
make_provider_config(
api_key="k",
base_url=DEEPSEEK_DEFAULT_BASE,
- rate_limit=1,
- rate_window=1,
),
admission=immediate_admission(),
)
@@ -689,8 +685,6 @@ def test_thinking_off_preserves_historical_reasoning():
make_provider_config(
api_key="k",
base_url=DEEPSEEK_DEFAULT_BASE,
- rate_limit=1,
- rate_window=1,
),
admission=immediate_admission(),
)
@@ -718,8 +712,6 @@ def test_thinking_off_still_replays_required_tool_reasoning():
make_provider_config(
api_key="k",
base_url=DEEPSEEK_DEFAULT_BASE,
- rate_limit=1,
- rate_window=1,
),
admission=immediate_admission(),
)
@@ -866,8 +858,6 @@ def test_vision_model_strips_user_document():
make_provider_config(
api_key="k",
base_url=DEEPSEEK_DEFAULT_BASE,
- rate_limit=1,
- rate_window=1,
),
admission=immediate_admission(),
)
@@ -903,8 +893,6 @@ def test_startup_rejects_mcp_servers():
make_provider_config(
api_key="k",
base_url=DEEPSEEK_DEFAULT_BASE,
- rate_limit=1,
- rate_window=1,
),
admission=immediate_admission(),
)
@@ -922,8 +910,6 @@ def test_startup_rejects_listed_server_tools_in_tools_list():
make_provider_config(
api_key="k",
base_url=DEEPSEEK_DEFAULT_BASE,
- rate_limit=1,
- rate_window=1,
),
admission=immediate_admission(),
)
@@ -959,8 +945,6 @@ def test_startup_preserves_completed_server_tool_history():
make_provider_config(
api_key="k",
base_url=DEEPSEEK_DEFAULT_BASE,
- rate_limit=1,
- rate_window=1,
),
admission=immediate_admission(),
)
diff --git a/tests/providers/test_execution_failure_boundary.py b/tests/providers/test_execution_failure_boundary.py
index 20d27a79..bdbdee26 100644
--- a/tests/providers/test_execution_failure_boundary.py
+++ b/tests/providers/test_execution_failure_boundary.py
@@ -62,8 +62,6 @@ def _provider() -> NvidiaNimProvider:
make_provider_config(
api_key="test_key",
base_url="https://test.api.nvidia.com/v1",
- rate_limit=10,
- rate_window=60,
),
nim_settings=NimSettings(),
admission=immediate_admission(),
diff --git a/tests/providers/test_experiential.py b/tests/providers/test_experiential.py
index f192276a..36f042bc 100644
--- a/tests/providers/test_experiential.py
+++ b/tests/providers/test_experiential.py
@@ -32,8 +32,6 @@ def experiential_provider() -> OpenAIChatProvider:
make_provider_config(
api_key="test-experiential-key",
base_url=EXPERIENTIAL_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(provider_name="experiential", max_attempts=1),
)
diff --git a/tests/providers/test_featherless.py b/tests/providers/test_featherless.py
index 4dd33ea0..eaf6b363 100644
--- a/tests/providers/test_featherless.py
+++ b/tests/providers/test_featherless.py
@@ -35,8 +35,6 @@ def featherless_provider() -> OpenAIChatProvider:
make_provider_config(
api_key="test-featherless-key",
base_url=FEATHERLESS_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(
provider_name="featherless",
diff --git a/tests/providers/test_fireworks.py b/tests/providers/test_fireworks.py
index 36f32ed9..7f33675c 100644
--- a/tests/providers/test_fireworks.py
+++ b/tests/providers/test_fireworks.py
@@ -25,8 +25,6 @@ def fireworks_provider():
make_provider_config(
api_key="test_fireworks_key",
base_url=FIREWORKS_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(),
)
@@ -81,8 +79,6 @@ def test_replay_is_independent_of_current_turn_reasoning_control():
make_provider_config(
api_key="k",
base_url=FIREWORKS_DEFAULT_BASE,
- rate_limit=1,
- rate_window=1,
),
admission=immediate_admission(),
)
diff --git a/tests/providers/test_gemini.py b/tests/providers/test_gemini.py
index cfa342d9..60b2c2d2 100644
--- a/tests/providers/test_gemini.py
+++ b/tests/providers/test_gemini.py
@@ -49,8 +49,6 @@ def gemini_config():
return make_provider_config(
api_key="test_gemini_key",
base_url=GEMINI_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
)
@@ -123,8 +121,6 @@ def test_build_request_body_reasoning_off_sets_reasoning_none():
make_provider_config(
api_key="test_gemini_key",
base_url=GEMINI_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(),
)
diff --git a/tests/providers/test_groq.py b/tests/providers/test_groq.py
index d329f56e..bc997e2d 100644
--- a/tests/providers/test_groq.py
+++ b/tests/providers/test_groq.py
@@ -24,8 +24,6 @@ def groq_config():
return make_provider_config(
api_key="test_groq_key",
base_url=GROQ_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
)
@@ -144,8 +142,6 @@ def test_build_request_body_global_disable_blocks_reasoning_mapping():
make_provider_config(
api_key="test_groq_key",
base_url=GROQ_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(),
)
diff --git a/tests/providers/test_groq_reasoning.py b/tests/providers/test_groq_reasoning.py
index 88909a6c..aaa22785 100644
--- a/tests/providers/test_groq_reasoning.py
+++ b/tests/providers/test_groq_reasoning.py
@@ -64,8 +64,6 @@ def _provider(*, max_attempts: int = 5) -> GroqProvider:
make_provider_config(
api_key="test_groq_key",
base_url=GROQ_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(
provider_name="GROQ",
diff --git a/tests/providers/test_huggingface.py b/tests/providers/test_huggingface.py
index f460d97a..e84f4822 100644
--- a/tests/providers/test_huggingface.py
+++ b/tests/providers/test_huggingface.py
@@ -29,8 +29,6 @@ def huggingface_config():
return make_provider_config(
api_key="test_hf_key",
base_url=HUGGINGFACE_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
)
diff --git a/tests/providers/test_kilo.py b/tests/providers/test_kilo.py
index 37fd73b0..b023ac48 100644
--- a/tests/providers/test_kilo.py
+++ b/tests/providers/test_kilo.py
@@ -34,8 +34,6 @@ def kilo_config():
return make_provider_config(
api_key="test_kilo_key",
base_url=KILO_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
)
diff --git a/tests/providers/test_kimi.py b/tests/providers/test_kimi.py
index 66cb8dde..aa88a43a 100644
--- a/tests/providers/test_kimi.py
+++ b/tests/providers/test_kimi.py
@@ -26,8 +26,6 @@ def kimi_provider():
make_provider_config(
api_key="test_kimi_key",
base_url=KIMI_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(),
)
diff --git a/tests/providers/test_kimi_code.py b/tests/providers/test_kimi_code.py
index 80e7aa22..445b0d57 100644
--- a/tests/providers/test_kimi_code.py
+++ b/tests/providers/test_kimi_code.py
@@ -34,8 +34,6 @@ def kimi_code_provider() -> OpenAIChatProvider:
make_provider_config(
api_key="test-subscription-key",
base_url=KIMI_CODE_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(),
)
diff --git a/tests/providers/test_lightning.py b/tests/providers/test_lightning.py
index 19451cc9..3f65391a 100644
--- a/tests/providers/test_lightning.py
+++ b/tests/providers/test_lightning.py
@@ -35,8 +35,6 @@ def lightning_provider() -> OpenAIChatProvider:
make_provider_config(
api_key="test-lightning-key",
base_url=LIGHTNING_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(provider_name="lightning", max_attempts=1),
)
diff --git a/tests/providers/test_llm7.py b/tests/providers/test_llm7.py
index ba871968..f7fbaed8 100644
--- a/tests/providers/test_llm7.py
+++ b/tests/providers/test_llm7.py
@@ -36,8 +36,6 @@ def llm7_provider() -> OpenAIChatProvider:
make_provider_config(
api_key="test-llm7-key",
base_url=LLM7_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(provider_name="llm7", max_attempts=1),
)
diff --git a/tests/providers/test_lmstudio.py b/tests/providers/test_lmstudio.py
index 4b762b6e..6d9de49f 100644
--- a/tests/providers/test_lmstudio.py
+++ b/tests/providers/test_lmstudio.py
@@ -49,8 +49,6 @@ def lmstudio_config():
return make_provider_config(
api_key="lm-studio",
base_url=LMSTUDIO_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
)
diff --git a/tests/providers/test_minimax.py b/tests/providers/test_minimax.py
index 21570795..a8a789ce 100644
--- a/tests/providers/test_minimax.py
+++ b/tests/providers/test_minimax.py
@@ -46,8 +46,6 @@ def minimax_provider():
make_provider_config(
api_key="test-minimax-key",
base_url=MINIMAX_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(),
)
diff --git a/tests/providers/test_mistral.py b/tests/providers/test_mistral.py
index e3b05443..1fce2eca 100644
--- a/tests/providers/test_mistral.py
+++ b/tests/providers/test_mistral.py
@@ -31,8 +31,6 @@ def mistral_config():
return make_provider_config(
api_key="test_mistral_key",
base_url=MISTRAL_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
)
@@ -209,8 +207,6 @@ def test_build_request_body_reasoning_off_uses_native_none():
make_provider_config(
api_key="test_mistral_key",
base_url=MISTRAL_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(),
)
@@ -226,8 +222,6 @@ def test_reasoning_off_keeps_replay_separate_from_new_turn_compute():
make_provider_config(
api_key="test_mistral_key",
base_url=MISTRAL_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(),
)
diff --git a/tests/providers/test_model_discovery.py b/tests/providers/test_model_discovery.py
index 3a656538..087d47c0 100644
--- a/tests/providers/test_model_discovery.py
+++ b/tests/providers/test_model_discovery.py
@@ -73,7 +73,9 @@ def _manager(
providers = providers or {}
return ProviderRuntimeManager(
settings,
- runtime_factory=lambda snapshot: ProviderRuntime(snapshot, dict(providers)),
+ runtime_factory=lambda snapshot, admission_registry: ProviderRuntime(
+ snapshot, admission_registry, dict(providers)
+ ),
)
diff --git a/tests/providers/test_nararoute.py b/tests/providers/test_nararoute.py
index b7d1b5b9..783c5a55 100644
--- a/tests/providers/test_nararoute.py
+++ b/tests/providers/test_nararoute.py
@@ -22,8 +22,6 @@ async def test_model_catalog_extracts_strict_optional_reasoning_boolean() -> Non
make_provider_config(
api_key="test-nararoute-key",
base_url=NARAROUTE_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(provider_name="nararoute"),
)
diff --git a/tests/providers/test_nebius.py b/tests/providers/test_nebius.py
index d50e0196..40f1d0d5 100644
--- a/tests/providers/test_nebius.py
+++ b/tests/providers/test_nebius.py
@@ -33,8 +33,6 @@ def nebius_provider() -> OpenAIChatProvider:
make_provider_config(
api_key="test-nebius-key",
base_url=NEBIUS_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(provider_name="nebius"),
)
diff --git a/tests/providers/test_novita.py b/tests/providers/test_novita.py
index 8b204366..8751d144 100644
--- a/tests/providers/test_novita.py
+++ b/tests/providers/test_novita.py
@@ -27,8 +27,6 @@ def novita_provider():
make_provider_config(
api_key="test_novita_key",
base_url=NOVITA_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(),
)
diff --git a/tests/providers/test_nvidia_nim_degraded_retry.py b/tests/providers/test_nvidia_nim_degraded_retry.py
index fef34494..2d2a68fb 100644
--- a/tests/providers/test_nvidia_nim_degraded_retry.py
+++ b/tests/providers/test_nvidia_nim_degraded_retry.py
@@ -31,9 +31,6 @@ def _config(base_url: str) -> ProviderConfig:
return make_provider_config(
api_key="test_key",
base_url=base_url,
- rate_limit=1_000_000,
- rate_window=1,
- max_concurrency=1_000,
http_read_timeout=30.0,
http_write_timeout=15.0,
http_connect_timeout=5.0,
diff --git a/tests/providers/test_open_router.py b/tests/providers/test_open_router.py
index e6ad1573..fcbfde7f 100644
--- a/tests/providers/test_open_router.py
+++ b/tests/providers/test_open_router.py
@@ -156,8 +156,6 @@ def open_router_provider():
make_provider_config(
api_key="test_openrouter_key",
base_url="https://openrouter.ai/api/v1",
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(),
)
diff --git a/tests/providers/test_openai_chat_output_cap.py b/tests/providers/test_openai_chat_output_cap.py
index c70e98bc..fbe28c15 100644
--- a/tests/providers/test_openai_chat_output_cap.py
+++ b/tests/providers/test_openai_chat_output_cap.py
@@ -233,8 +233,6 @@ def groq_provider():
make_provider_config(
api_key="test_groq_key",
base_url=GROQ_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(),
)
diff --git a/tests/providers/test_openai_chat_usage.py b/tests/providers/test_openai_chat_usage.py
index e0660e49..8ab19d2d 100644
--- a/tests/providers/test_openai_chat_usage.py
+++ b/tests/providers/test_openai_chat_usage.py
@@ -56,8 +56,6 @@ class _UsageTestProvider(OpenAIChatProvider):
make_provider_config(
api_key="test_key",
base_url="https://provider.example/v1",
- rate_limit=100,
- rate_window=60,
),
behavior=_UsageTestBehavior(
OpenAIChatProfile(
diff --git a/tests/providers/test_openai_codex_provider.py b/tests/providers/test_openai_codex_provider.py
index 87f1801e..9dfdddb7 100644
--- a/tests/providers/test_openai_codex_provider.py
+++ b/tests/providers/test_openai_codex_provider.py
@@ -92,9 +92,6 @@ def _config() -> ProviderConfig:
return make_provider_config(
api_key="",
base_url="https://chatgpt.com/backend-api/codex",
- rate_limit=100,
- rate_window=1,
- max_concurrency=2,
)
diff --git a/tests/providers/test_openai_compat_5xx_retry.py b/tests/providers/test_openai_compat_5xx_retry.py
index b848335b..bee395b4 100644
--- a/tests/providers/test_openai_compat_5xx_retry.py
+++ b/tests/providers/test_openai_compat_5xx_retry.py
@@ -40,8 +40,6 @@ async def test_nim_stream_retries_on_openai_5xx_then_streams(status_code):
config = make_provider_config(
api_key="test_key",
base_url="https://test.api.nvidia.com/v1",
- rate_limit=100,
- rate_window=60,
http_read_timeout=600.0,
http_write_timeout=15.0,
http_connect_timeout=5.0,
@@ -87,8 +85,6 @@ async def test_nim_stream_retries_on_pre_stream_connection_error_then_streams():
config = make_provider_config(
api_key="test_key",
base_url="https://test.api.nvidia.com/v1",
- rate_limit=100,
- rate_window=60,
http_read_timeout=600.0,
http_write_timeout=15.0,
http_connect_timeout=5.0,
@@ -131,8 +127,6 @@ async def test_nim_stream_connection_error_exhausted_emits_cause_chain():
config = make_provider_config(
api_key="test_key",
base_url="https://test.api.nvidia.com/v1",
- rate_limit=100,
- rate_window=60,
http_read_timeout=600.0,
http_write_timeout=15.0,
http_connect_timeout=5.0,
@@ -186,8 +180,6 @@ async def test_nim_stream_openai_5xx_exhausted_emits_user_message(
config = make_provider_config(
api_key="test_key",
base_url="https://test.api.nvidia.com/v1",
- rate_limit=100,
- rate_window=60,
http_read_timeout=600.0,
http_write_timeout=15.0,
http_connect_timeout=5.0,
diff --git a/tests/providers/test_opencode.py b/tests/providers/test_opencode.py
index 13040dcf..aea6338c 100644
--- a/tests/providers/test_opencode.py
+++ b/tests/providers/test_opencode.py
@@ -58,8 +58,6 @@ def _config():
return make_provider_config(
api_key="test_opencode_key",
base_url="https://opencode.ai/zen/v1",
- rate_limit=100,
- rate_window=1,
)
diff --git a/tests/providers/test_orcarouter.py b/tests/providers/test_orcarouter.py
index 1ee38f95..b67539e0 100644
--- a/tests/providers/test_orcarouter.py
+++ b/tests/providers/test_orcarouter.py
@@ -35,8 +35,6 @@ def orcarouter_provider() -> OpenAIChatProvider:
make_provider_config(
api_key="test-orcarouter-key",
base_url=ORCAROUTER_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(provider_name="orcarouter", max_attempts=1),
)
diff --git a/tests/providers/test_poolside.py b/tests/providers/test_poolside.py
index 49eadc81..efa7422a 100644
--- a/tests/providers/test_poolside.py
+++ b/tests/providers/test_poolside.py
@@ -34,8 +34,6 @@ def poolside_provider() -> OpenAIChatProvider:
make_provider_config(
api_key="test-poolside-key",
base_url=POOLSIDE_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(provider_name="poolside", max_attempts=1),
)
diff --git a/tests/providers/test_provider_client_tool_discovery.py b/tests/providers/test_provider_client_tool_discovery.py
index f4bc940f..0910c635 100644
--- a/tests/providers/test_provider_client_tool_discovery.py
+++ b/tests/providers/test_provider_client_tool_discovery.py
@@ -9,6 +9,8 @@ from openai import AsyncOpenAI
from free_claude_code.config.settings import Settings
from free_claude_code.core.anthropic.stream_contracts import parse_sse_text
from free_claude_code.core.openai_responses import OpenAIResponsesRequest
+from free_claude_code.providers.admission_policy import ProviderAdmissionLimits
+from free_claude_code.providers.admission_registry import ProviderAdmissionRegistry
from free_claude_code.providers.openai_chat import OpenAIChatProvider
from free_claude_code.providers.runtime.runtime import create_provider
from tests.core.openai_responses.test_client_tool_discovery import AGENTS, SEARCH
@@ -257,6 +259,16 @@ async def test_provider_discovery_call_and_result_round_trip(
groq_api_key="test",
mistral_api_key="test",
),
+ ProviderAdmissionRegistry(
+ ProviderAdmissionLimits.from_settings(
+ Settings(
+ open_router_api_key="test",
+ nvidia_nim_api_key="test",
+ groq_api_key="test",
+ mistral_api_key="test",
+ )
+ )
+ ),
)
),
)
diff --git a/tests/providers/test_provider_runtime.py b/tests/providers/test_provider_runtime.py
index 02ecce18..da9a216f 100644
--- a/tests/providers/test_provider_runtime.py
+++ b/tests/providers/test_provider_runtime.py
@@ -44,6 +44,8 @@ from free_claude_code.config.provider_catalog import (
ZENMUX_DEFAULT_BASE,
)
from free_claude_code.providers.admission import ProviderAdmissionController
+from free_claude_code.providers.admission_policy import ProviderAdmissionLimits
+from free_claude_code.providers.admission_registry import ProviderAdmissionRegistry
from free_claude_code.providers.cloudflare import CloudflareProvider
from free_claude_code.providers.deepseek import DeepSeekProvider
from free_claude_code.providers.gemini import GeminiProvider
@@ -261,7 +263,11 @@ async def test_poolside_provider_config_uses_key_base_and_proxy() -> None:
config = build_provider_config(descriptor, settings)
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
- provider = await create_provider("poolside", settings)
+ provider = await create_provider(
+ "poolside",
+ settings,
+ ProviderAdmissionRegistry(ProviderAdmissionLimits.from_settings(settings)),
+ )
assert descriptor.display_name == "Poolside AI"
assert descriptor.credential_env == "POOLSIDE_API_KEY"
@@ -286,7 +292,11 @@ async def test_llm7_provider_config_uses_key_base_and_proxy() -> None:
config = build_provider_config(descriptor, settings)
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
- provider = await create_provider("llm7", settings)
+ provider = await create_provider(
+ "llm7",
+ settings,
+ ProviderAdmissionRegistry(ProviderAdmissionLimits.from_settings(settings)),
+ )
assert descriptor.display_name == "LLM7.io"
assert descriptor.credential_env == "LLM7_API_KEY"
@@ -311,7 +321,11 @@ async def test_experiential_provider_config_uses_key_base_and_proxy() -> None:
config = build_provider_config(descriptor, settings)
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
- provider = await create_provider("experiential", settings)
+ provider = await create_provider(
+ "experiential",
+ settings,
+ ProviderAdmissionRegistry(ProviderAdmissionLimits.from_settings(settings)),
+ )
assert descriptor.display_name == "Experiential Labs"
assert descriptor.credential_env == "EXPLABS_API_KEY"
@@ -339,7 +353,11 @@ async def test_cheaperinference_provider_config_uses_key_base_and_proxy() -> Non
config = build_provider_config(descriptor, settings)
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
- provider = await create_provider("cheaperinference", settings)
+ provider = await create_provider(
+ "cheaperinference",
+ settings,
+ ProviderAdmissionRegistry(ProviderAdmissionLimits.from_settings(settings)),
+ )
assert descriptor.display_name == "Cheaper Inference"
assert descriptor.credential_env == "CHEAPER_INFERENCE_API_KEY"
@@ -365,7 +383,11 @@ async def test_orcarouter_provider_config_uses_key_base_and_proxy() -> None:
config = build_provider_config(descriptor, settings)
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
- provider = await create_provider("orcarouter", settings)
+ provider = await create_provider(
+ "orcarouter",
+ settings,
+ ProviderAdmissionRegistry(ProviderAdmissionLimits.from_settings(settings)),
+ )
assert descriptor.display_name == "OrcaRouter"
assert descriptor.credential_env == "ORCAROUTER_API_KEY"
@@ -390,7 +412,11 @@ async def test_xai_provider_config_uses_key_base_and_proxy() -> None:
config = build_provider_config(descriptor, settings)
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
- provider = await create_provider("xai", settings)
+ provider = await create_provider(
+ "xai",
+ settings,
+ ProviderAdmissionRegistry(ProviderAdmissionLimits.from_settings(settings)),
+ )
assert descriptor.display_name == "xAI (Grok)"
assert descriptor.credential_env == "XAI_API_KEY"
@@ -410,7 +436,11 @@ async def test_qwencloud_provider_config_uses_key_base_and_proxy() -> None:
config = build_provider_config(descriptor, settings)
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
- provider = await create_provider("qwencloud", settings)
+ provider = await create_provider(
+ "qwencloud",
+ settings,
+ ProviderAdmissionRegistry(ProviderAdmissionLimits.from_settings(settings)),
+ )
assert descriptor.display_name == "QwenCloud Token Plan"
assert descriptor.credential_env == "QWENCLOUD_API_KEY"
@@ -430,7 +460,11 @@ async def test_cline_pass_provider_config_uses_key_base_and_proxy() -> None:
config = build_provider_config(descriptor, settings)
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
- provider = await create_provider("cline_pass", settings)
+ provider = await create_provider(
+ "cline_pass",
+ settings,
+ ProviderAdmissionRegistry(ProviderAdmissionLimits.from_settings(settings)),
+ )
assert descriptor.display_name == "ClinePass"
assert descriptor.credential_env == "CLINE_API_KEY"
@@ -454,7 +488,11 @@ async def test_qwencloud_coding_provider_config_uses_key_base_and_proxy() -> Non
config = build_provider_config(descriptor, settings)
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
- provider = await create_provider("qwencloud_coding", settings)
+ provider = await create_provider(
+ "qwencloud_coding",
+ settings,
+ ProviderAdmissionRegistry(ProviderAdmissionLimits.from_settings(settings)),
+ )
assert descriptor.display_name == "QwenCloud Coding Plan"
assert descriptor.credential_env == "QWENCLOUD_CODING_API_KEY"
@@ -474,7 +512,11 @@ async def test_together_provider_config_uses_key_base_and_proxy() -> None:
config = build_provider_config(descriptor, settings)
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
- provider = await create_provider("together", settings)
+ provider = await create_provider(
+ "together",
+ settings,
+ ProviderAdmissionRegistry(ProviderAdmissionLimits.from_settings(settings)),
+ )
assert descriptor.display_name == "Together AI"
assert descriptor.credential_env == "TOGETHER_API_KEY"
@@ -494,7 +536,11 @@ async def test_deepinfra_provider_config_uses_key_base_and_proxy() -> None:
config = build_provider_config(descriptor, settings)
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
- provider = await create_provider("deepinfra", settings)
+ provider = await create_provider(
+ "deepinfra",
+ settings,
+ ProviderAdmissionRegistry(ProviderAdmissionLimits.from_settings(settings)),
+ )
assert descriptor.display_name == "DeepInfra"
assert descriptor.credential_env == "DEEPINFRA_API_KEY"
@@ -514,7 +560,11 @@ async def test_siliconflow_provider_config_uses_key_base_and_proxy() -> None:
config = build_provider_config(descriptor, settings)
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
- provider = await create_provider("siliconflow", settings)
+ provider = await create_provider(
+ "siliconflow",
+ settings,
+ ProviderAdmissionRegistry(ProviderAdmissionLimits.from_settings(settings)),
+ )
assert descriptor.display_name == "SiliconFlow"
assert descriptor.credential_env == "SILICONFLOW_API_KEY"
@@ -535,7 +585,11 @@ async def test_nebius_provider_config_uses_key_base_and_proxy() -> None:
config = build_provider_config(descriptor, settings)
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
- provider = await create_provider("nebius", settings)
+ provider = await create_provider(
+ "nebius",
+ settings,
+ ProviderAdmissionRegistry(ProviderAdmissionLimits.from_settings(settings)),
+ )
assert descriptor.display_name == "Nebius Token Factory"
assert descriptor.credential_env == "NEBIUS_API_KEY"
@@ -559,7 +613,11 @@ async def test_scaleway_provider_config_uses_key_base_and_proxy() -> None:
config = build_provider_config(descriptor, settings)
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
- provider = await create_provider("scaleway", settings)
+ provider = await create_provider(
+ "scaleway",
+ settings,
+ ProviderAdmissionRegistry(ProviderAdmissionLimits.from_settings(settings)),
+ )
assert descriptor.display_name == "Scaleway"
assert descriptor.credential_env == "SCW_SECRET_KEY"
@@ -581,7 +639,11 @@ async def test_chutes_provider_config_uses_key_base_and_proxy() -> None:
config = build_provider_config(descriptor, settings)
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
- provider = await create_provider("chutes", settings)
+ provider = await create_provider(
+ "chutes",
+ settings,
+ ProviderAdmissionRegistry(ProviderAdmissionLimits.from_settings(settings)),
+ )
assert descriptor.display_name == "Chutes"
assert descriptor.credential_env == "CHUTES_API_KEY"
@@ -605,7 +667,11 @@ async def test_featherless_provider_config_uses_key_base_and_proxy() -> None:
config = build_provider_config(descriptor, settings)
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
- provider = await create_provider("featherless", settings)
+ provider = await create_provider(
+ "featherless",
+ settings,
+ ProviderAdmissionRegistry(ProviderAdmissionLimits.from_settings(settings)),
+ )
assert descriptor.display_name == "Featherless AI"
assert descriptor.credential_env == "FEATHERLESS_API_KEY"
@@ -627,7 +693,11 @@ async def test_agnes_provider_config_uses_key_base_and_proxy() -> None:
config = build_provider_config(descriptor, settings)
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
- provider = await create_provider("agnes", settings)
+ provider = await create_provider(
+ "agnes",
+ settings,
+ ProviderAdmissionRegistry(ProviderAdmissionLimits.from_settings(settings)),
+ )
assert descriptor.display_name == "Agnes AI"
assert descriptor.credential_env == "AGNES_API_KEY"
@@ -648,7 +718,11 @@ async def test_zenmux_provider_config_uses_key_base_and_proxy() -> None:
config = build_provider_config(descriptor, settings)
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
- provider = await create_provider("zenmux", settings)
+ provider = await create_provider(
+ "zenmux",
+ settings,
+ ProviderAdmissionRegistry(ProviderAdmissionLimits.from_settings(settings)),
+ )
assert descriptor.display_name == "ZenMux"
assert descriptor.credential_env == "ZENMUX_API_KEY"
@@ -669,7 +743,11 @@ async def test_wandb_provider_config_uses_key_base_and_proxy() -> None:
config = build_provider_config(descriptor, settings)
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
- provider = await create_provider("wandb", settings)
+ provider = await create_provider(
+ "wandb",
+ settings,
+ ProviderAdmissionRegistry(ProviderAdmissionLimits.from_settings(settings)),
+ )
assert descriptor.display_name == "W&B Inference"
assert descriptor.credential_env == "WANDB_API_KEY"
@@ -752,7 +830,11 @@ async def test_local_provider_factory_resolves_catalog_static_credential(
config = build_provider_config(descriptor, settings)
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
- provider = await create_provider(provider_id, settings)
+ provider = await create_provider(
+ provider_id,
+ settings,
+ ProviderAdmissionRegistry(ProviderAdmissionLimits.from_settings(settings)),
+ )
assert config.api_key == expected_api_key
assert isinstance(provider, OpenAIChatProvider)
@@ -788,7 +870,11 @@ async def test_zai_api_provider_config_uses_shared_key_general_base_and_own_prox
config = build_provider_config(descriptor, settings)
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
- provider = await create_provider("zai_api", settings)
+ provider = await create_provider(
+ "zai_api",
+ settings,
+ ProviderAdmissionRegistry(ProviderAdmissionLimits.from_settings(settings)),
+ )
assert descriptor.display_name == "Z.ai API"
assert descriptor.credential_env == "ZAI_API_KEY"
@@ -840,7 +926,11 @@ async def test_create_cloudflare_provider_uses_account_scoped_base_url():
)
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
- provider = await create_provider("cloudflare", settings)
+ provider = await create_provider(
+ "cloudflare",
+ settings,
+ ProviderAdmissionRegistry(ProviderAdmissionLimits.from_settings(settings)),
+ )
assert isinstance(provider, CloudflareProvider)
assert provider._base_url == (
@@ -851,7 +941,13 @@ async def test_create_cloudflare_provider_uses_account_scoped_base_url():
@pytest.mark.asyncio
async def test_opencode_zen_provider_config_uses_explicit_id_and_name():
with patch("httpx.AsyncClient"):
- provider = await create_provider("opencode_zen", _make_settings())
+ provider = await create_provider(
+ "opencode_zen",
+ _make_settings(),
+ ProviderAdmissionRegistry(
+ ProviderAdmissionLimits.from_settings(_make_settings())
+ ),
+ )
assert isinstance(provider, OpenCodeProvider)
assert str(provider._client.base_url).rstrip("/") == "https://opencode.ai/zen/v1"
@@ -863,7 +959,13 @@ async def test_opencode_zen_provider_config_uses_explicit_id_and_name():
@pytest.mark.asyncio
async def test_opencode_go_provider_config_uses_correct_base_url_and_name():
with patch("httpx.AsyncClient"):
- provider = await create_provider("opencode_go", _make_settings())
+ provider = await create_provider(
+ "opencode_go",
+ _make_settings(),
+ ProviderAdmissionRegistry(
+ ProviderAdmissionLimits.from_settings(_make_settings())
+ ),
+ )
assert isinstance(provider, OpenCodeProvider)
assert str(provider._client.base_url).rstrip("/") == "https://opencode.ai/zen/go/v1"
@@ -954,7 +1056,13 @@ def test_build_provider_config_cohere_uses_api_key_and_proxy() -> None:
@pytest.mark.asyncio
async def test_create_provider_uses_openai_chat_openrouter_by_default():
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
- provider = await create_provider("open_router", _make_settings())
+ provider = await create_provider(
+ "open_router",
+ _make_settings(),
+ ProviderAdmissionRegistry(
+ ProviderAdmissionLimits.from_settings(_make_settings())
+ ),
+ )
assert isinstance(provider, OpenRouterProvider)
@@ -1062,7 +1170,7 @@ async def test_create_provider_instantiates_each_builtin():
patch("free_claude_code.providers.openai_api.provider.AsyncOpenAI"),
patch("httpx.AsyncClient"),
patch(
- "free_claude_code.providers.runtime.factory.ProviderAdmissionController",
+ "free_claude_code.providers.admission.ProviderAdmissionController",
return_value=sentinel_admission,
) as admission_factory,
):
@@ -1070,6 +1178,9 @@ async def test_create_provider_instantiates_each_builtin():
provider = await create_provider(
provider_id,
settings,
+ ProviderAdmissionRegistry(
+ ProviderAdmissionLimits.from_settings(settings)
+ ),
provider_loaders={
key: lambda factory=factory: factory
for key, factory in injected_factories.items()
@@ -1095,7 +1206,12 @@ async def test_create_provider_instantiates_each_builtin():
@pytest.mark.asyncio
async def test_provider_runtime_caches_by_provider_id():
- runtime = ProviderRuntime(_make_settings())
+ runtime = ProviderRuntime(
+ _make_settings(),
+ ProviderAdmissionRegistry(
+ ProviderAdmissionLimits.from_settings(_make_settings())
+ ),
+ )
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
first = await runtime.resolve_provider("nvidia_nim")
@@ -1111,7 +1227,7 @@ async def test_provider_creation_retries_after_shared_failure(cancelled):
provider = MagicMock(cleanup=AsyncMock())
attempts = 0
- async def construct(_id, _settings):
+ async def construct(_id, _settings, _admission_registry):
nonlocal attempts
attempts += 1
attempt = attempts
@@ -1123,7 +1239,13 @@ async def test_provider_creation_retries_after_shared_failure(cancelled):
raise RuntimeError("construction failed")
return provider
- runtime = ProviderRuntime(_make_settings(), provider_constructor=construct)
+ runtime = ProviderRuntime(
+ _make_settings(),
+ ProviderAdmissionRegistry(
+ ProviderAdmissionLimits.from_settings(_make_settings())
+ ),
+ provider_constructor=construct,
+ )
waiting = []
try:
for attempt in (1, 2):
@@ -1161,7 +1283,7 @@ async def test_finished_creation_callback_cannot_forget_a_new_retry():
attempts = 0
retry = None
- async def construct(_id, _settings):
+ async def construct(_id, _settings, _admission_registry):
nonlocal attempts, retry
attempts += 1
if attempts == 1:
@@ -1172,7 +1294,13 @@ async def test_finished_creation_callback_cannot_forget_a_new_retry():
await release.wait()
return provider
- runtime = ProviderRuntime(_make_settings(), provider_constructor=construct)
+ runtime = ProviderRuntime(
+ _make_settings(),
+ ProviderAdmissionRegistry(
+ ProviderAdmissionLimits.from_settings(_make_settings())
+ ),
+ provider_constructor=construct,
+ )
later = None
try:
with pytest.raises(RuntimeError, match="construction failed"):
@@ -1200,14 +1328,20 @@ async def test_cancelled_acquisition_keeps_construction_available_to_next_caller
provider = MagicMock(cleanup=AsyncMock())
attempts = 0
- async def construct(_id, _settings):
+ async def construct(_id, _settings, _admission_registry):
nonlocal attempts
attempts += 1
entered.set()
await release.wait()
return provider
- runtime = ProviderRuntime(_make_settings(), provider_constructor=construct)
+ runtime = ProviderRuntime(
+ _make_settings(),
+ ProviderAdmissionRegistry(
+ ProviderAdmissionLimits.from_settings(_make_settings())
+ ),
+ provider_constructor=construct,
+ )
first = asyncio.create_task(runtime.resolve_provider("nvidia_nim"))
second = None
try:
@@ -1236,13 +1370,19 @@ async def test_abandoned_creation_failure_has_no_unretrieved_exception():
previous_handler = loop.get_exception_handler()
errors = []
- async def construct(_id, _settings):
+ async def construct(_id, _settings, _admission_registry):
entered.set()
await release.wait()
failed.set()
raise RuntimeError("abandoned construction failed")
- runtime = ProviderRuntime(_make_settings(), provider_constructor=construct)
+ runtime = ProviderRuntime(
+ _make_settings(),
+ ProviderAdmissionRegistry(
+ ProviderAdmissionLimits.from_settings(_make_settings())
+ ),
+ provider_constructor=construct,
+ )
waiting = asyncio.create_task(runtime.resolve_provider("nvidia_nim"))
loop.set_exception_handler(lambda _loop, context: errors.append(context))
try:
@@ -1272,14 +1412,20 @@ async def test_shutdown_drains_construction_that_finishes_during_cancellation():
entered = asyncio.Event()
provider = MagicMock(cleanup=AsyncMock())
- async def construct(_id, _settings):
+ async def construct(_id, _settings, _admission_registry):
entered.set()
try:
await asyncio.Event().wait()
except asyncio.CancelledError:
return provider
- runtime = ProviderRuntime(_make_settings(), provider_constructor=construct)
+ runtime = ProviderRuntime(
+ _make_settings(),
+ ProviderAdmissionRegistry(
+ ProviderAdmissionLimits.from_settings(_make_settings())
+ ),
+ provider_constructor=construct,
+ )
waiting = asyncio.create_task(runtime.resolve_provider("nvidia_nim"))
try:
await entered.wait()
@@ -1297,7 +1443,12 @@ async def test_shutdown_drains_construction_that_finishes_during_cancellation():
@pytest.mark.asyncio
async def test_provider_runtime_provider_owns_one_admission_controller() -> None:
- runtime = ProviderRuntime(_make_settings())
+ runtime = ProviderRuntime(
+ _make_settings(),
+ ProviderAdmissionRegistry(
+ ProviderAdmissionLimits.from_settings(_make_settings())
+ ),
+ )
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
first = await runtime.resolve_provider("nvidia_nim")
@@ -1310,8 +1461,18 @@ async def test_provider_runtime_provider_owns_one_admission_controller() -> None
@pytest.mark.asyncio
async def test_separate_provider_runtimes_never_share_admission_controllers() -> None:
- first_runtime = ProviderRuntime(_make_settings())
- second_runtime = ProviderRuntime(_make_settings())
+ first_runtime = ProviderRuntime(
+ _make_settings(),
+ ProviderAdmissionRegistry(
+ ProviderAdmissionLimits.from_settings(_make_settings())
+ ),
+ )
+ second_runtime = ProviderRuntime(
+ _make_settings(),
+ ProviderAdmissionRegistry(
+ ProviderAdmissionLimits.from_settings(_make_settings())
+ ),
+ )
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
first = await first_runtime.resolve_provider("nvidia_nim")
@@ -1325,7 +1486,12 @@ async def test_separate_provider_runtimes_never_share_admission_controllers() ->
@pytest.mark.asyncio
async def test_different_providers_have_independent_admission_controllers() -> None:
- runtime = ProviderRuntime(_make_settings())
+ runtime = ProviderRuntime(
+ _make_settings(),
+ ProviderAdmissionRegistry(
+ ProviderAdmissionLimits.from_settings(_make_settings())
+ ),
+ )
with patch("free_claude_code.providers.openai_chat.client.AsyncOpenAI"):
nim = await runtime.resolve_provider("nvidia_nim")
@@ -1339,7 +1505,15 @@ async def test_different_providers_have_independent_admission_controllers() -> N
@pytest.mark.asyncio
async def test_unknown_provider_raises_unknown_provider_type_error():
with pytest.raises(UnknownProviderError, match="Unknown provider_type"):
- (await create_provider("unknown", _make_settings()))
+ (
+ await create_provider(
+ "unknown",
+ _make_settings(),
+ ProviderAdmissionRegistry(
+ ProviderAdmissionLimits.from_settings(_make_settings())
+ ),
+ )
+ )
@pytest.mark.asyncio
@@ -1349,7 +1523,13 @@ async def test_provider_runtime_cleanup_runs_all_even_if_one_fails() -> None:
p1.cleanup = AsyncMock(side_effect=RuntimeError("first"))
p2 = MagicMock()
p2.cleanup = AsyncMock()
- runtime = ProviderRuntime(_make_settings(), {"a": p1, "b": p2})
+ runtime = ProviderRuntime(
+ _make_settings(),
+ ProviderAdmissionRegistry(
+ ProviderAdmissionLimits.from_settings(_make_settings())
+ ),
+ {"a": p1, "b": p2},
+ )
with pytest.raises(RuntimeError, match="first"):
await runtime.cleanup()
@@ -1386,6 +1566,9 @@ async def test_cancelled_cleanup_retains_current_and_unvisited_providers() -> No
third.cleanup = AsyncMock()
runtime = ProviderRuntime(
_make_settings(),
+ ProviderAdmissionRegistry(
+ ProviderAdmissionLimits.from_settings(_make_settings())
+ ),
{"first": first, "second": second, "third": third},
)
cleanup_task = asyncio.create_task(runtime.cleanup())
@@ -1417,7 +1600,13 @@ async def test_provider_runtime_cleanup_exceptiongroup_on_multiple_failures() ->
p1.cleanup = AsyncMock(side_effect=RuntimeError("a"))
p2 = MagicMock()
p2.cleanup = AsyncMock(side_effect=RuntimeError("b"))
- runtime = ProviderRuntime(_make_settings(), {"x": p1, "y": p2})
+ runtime = ProviderRuntime(
+ _make_settings(),
+ ProviderAdmissionRegistry(
+ ProviderAdmissionLimits.from_settings(_make_settings())
+ ),
+ {"x": p1, "y": p2},
+ )
with pytest.raises(ExceptionGroup) as exc_info:
await runtime.cleanup()
diff --git a/tests/providers/test_qwencloud.py b/tests/providers/test_qwencloud.py
index 03738553..8e6a1e5e 100644
--- a/tests/providers/test_qwencloud.py
+++ b/tests/providers/test_qwencloud.py
@@ -30,8 +30,6 @@ def qwencloud_provider() -> OpenAIChatProvider:
make_provider_config(
api_key="test-qwencloud-key",
base_url=QWENCLOUD_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(provider_name="qwencloud"),
)
diff --git a/tests/providers/test_qwencloud_coding.py b/tests/providers/test_qwencloud_coding.py
index 7a3be809..fca1b05a 100644
--- a/tests/providers/test_qwencloud_coding.py
+++ b/tests/providers/test_qwencloud_coding.py
@@ -28,8 +28,6 @@ def qwencloud_coding_provider() -> OpenAIChatProvider:
make_provider_config(
api_key="test-qwencloud-coding-key",
base_url=QWENCLOUD_CODING_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(provider_name="qwencloud_coding"),
)
diff --git a/tests/providers/test_sambanova.py b/tests/providers/test_sambanova.py
index dc82b219..cf5d5678 100644
--- a/tests/providers/test_sambanova.py
+++ b/tests/providers/test_sambanova.py
@@ -25,8 +25,6 @@ def sambanova_config():
return make_provider_config(
api_key="test_sambanova_key",
base_url=SAMBANOVA_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
)
diff --git a/tests/providers/test_scaleway.py b/tests/providers/test_scaleway.py
index 96969b16..5d38e640 100644
--- a/tests/providers/test_scaleway.py
+++ b/tests/providers/test_scaleway.py
@@ -33,8 +33,6 @@ def scaleway_provider() -> OpenAIChatProvider:
make_provider_config(
api_key="test-scaleway-key",
base_url=SCALEWAY_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(provider_name="scaleway"),
)
diff --git a/tests/providers/test_siliconflow.py b/tests/providers/test_siliconflow.py
index ee1ec0d9..7b46ed4d 100644
--- a/tests/providers/test_siliconflow.py
+++ b/tests/providers/test_siliconflow.py
@@ -45,8 +45,6 @@ def siliconflow_provider() -> OpenAIChatProvider:
make_provider_config(
api_key="test-siliconflow-key",
base_url=SILICONFLOW_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(provider_name="siliconflow"),
)
@@ -235,8 +233,6 @@ async def test_model_catalog_uses_documented_endpoint_query_and_auth() -> None:
make_provider_config(
api_key="wire-siliconflow-key",
base_url=SILICONFLOW_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(provider_name="siliconflow"),
)
diff --git a/tests/providers/test_streaming_errors.py b/tests/providers/test_streaming_errors.py
index 9c19e95f..6b05a5d8 100644
--- a/tests/providers/test_streaming_errors.py
+++ b/tests/providers/test_streaming_errors.py
@@ -127,8 +127,6 @@ def _make_provider():
config = make_provider_config(
api_key="test_key",
base_url="https://test.api.nvidia.com/v1",
- rate_limit=10,
- rate_window=60,
)
return NvidiaNimProvider(
config,
diff --git a/tests/providers/test_together.py b/tests/providers/test_together.py
index 7bd1bc47..7dcceb79 100644
--- a/tests/providers/test_together.py
+++ b/tests/providers/test_together.py
@@ -31,8 +31,6 @@ def together_provider() -> OpenAIChatProvider:
make_provider_config(
api_key="test-together-key",
base_url=TOGETHER_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(provider_name="together"),
)
diff --git a/tests/providers/test_vercel.py b/tests/providers/test_vercel.py
index daaa3883..8ba68f49 100644
--- a/tests/providers/test_vercel.py
+++ b/tests/providers/test_vercel.py
@@ -27,8 +27,6 @@ def vercel_config():
return make_provider_config(
api_key="test_vercel_key",
base_url=VERCEL_AI_GATEWAY_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
)
diff --git a/tests/providers/test_wafer.py b/tests/providers/test_wafer.py
index 1172c726..d46aeb02 100644
--- a/tests/providers/test_wafer.py
+++ b/tests/providers/test_wafer.py
@@ -25,8 +25,6 @@ def wafer_config():
return make_provider_config(
api_key="test-wafer-key",
base_url=WAFER_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
)
diff --git a/tests/providers/test_wandb.py b/tests/providers/test_wandb.py
index 0184f68e..9cf0d665 100644
--- a/tests/providers/test_wandb.py
+++ b/tests/providers/test_wandb.py
@@ -33,8 +33,6 @@ def wandb_provider() -> OpenAIChatProvider:
make_provider_config(
api_key="test-wandb-key",
base_url=WANDB_INFERENCE_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(provider_name="wandb"),
)
@@ -209,8 +207,6 @@ async def test_model_catalog_uses_documented_endpoint_and_bearer_auth() -> None:
make_provider_config(
api_key="wire-wandb-key",
base_url=WANDB_INFERENCE_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(provider_name="wandb"),
)
diff --git a/tests/providers/test_xai.py b/tests/providers/test_xai.py
index f023d49c..5fe601ed 100644
--- a/tests/providers/test_xai.py
+++ b/tests/providers/test_xai.py
@@ -32,8 +32,6 @@ def xai_provider() -> OpenAIChatProvider:
make_provider_config(
api_key="test-xai-key",
base_url=XAI_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(provider_name="xai"),
)
diff --git a/tests/providers/test_zai.py b/tests/providers/test_zai.py
index f9f3f7e3..25dfeb8b 100644
--- a/tests/providers/test_zai.py
+++ b/tests/providers/test_zai.py
@@ -33,8 +33,6 @@ def zai_provider(request: pytest.FixtureRequest):
make_provider_config(
api_key="test_zai_key",
base_url=_ZAI_BASES[provider_id],
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(),
)
diff --git a/tests/providers/test_zenmux.py b/tests/providers/test_zenmux.py
index 91cd6031..fcf7b2ea 100644
--- a/tests/providers/test_zenmux.py
+++ b/tests/providers/test_zenmux.py
@@ -43,8 +43,6 @@ def zenmux_provider() -> OpenAIChatProvider:
make_provider_config(
api_key="test-zenmux-key",
base_url=ZENMUX_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(provider_name="zenmux"),
)
@@ -456,8 +454,6 @@ async def test_model_catalog_uses_documented_endpoint_and_auth() -> None:
make_provider_config(
api_key="wire-zenmux-key",
base_url=ZENMUX_DEFAULT_BASE,
- rate_limit=10,
- rate_window=60,
),
admission=immediate_admission(provider_name="zenmux"),
)
diff --git a/tests/runtime/test_application_runtime.py b/tests/runtime/test_application_runtime.py
index 1feb66d1..7ee11375 100644
--- a/tests/runtime/test_application_runtime.py
+++ b/tests/runtime/test_application_runtime.py
@@ -37,8 +37,8 @@ from tests.providers.support import make_provider_config
class TrackingRuntime(ProviderRuntime):
- def __init__(self, settings: Settings) -> None:
- super().__init__(settings)
+ def __init__(self, settings: Settings, admission_registry) -> None:
+ super().__init__(settings, admission_registry)
self.cleanup_calls = 0
async def cleanup(self) -> None:
@@ -105,11 +105,11 @@ class TrackingFactory:
self.fail = False
self.events: list[str] = []
- def __call__(self, settings: Settings) -> ProviderRuntime:
+ def __call__(self, settings: Settings, admission_registry) -> ProviderRuntime:
self.events.append(f"construct:{settings.model}")
if self.fail:
raise RuntimeError("candidate failed")
- runtime = TrackingRuntime(settings)
+ runtime = TrackingRuntime(settings, admission_registry)
self.runtimes.append(runtime)
return runtime
@@ -233,8 +233,9 @@ def _runtime_with_admin_provider(
)
manager = ProviderRuntimeManager(
settings,
- runtime_factory=lambda snapshot: ProviderRuntime(
+ runtime_factory=lambda snapshot, admission_registry: ProviderRuntime(
snapshot,
+ admission_registry,
{"nvidia_nim": provider},
),
)
diff --git a/tests/runtime/test_custom_provider_lifecycle.py b/tests/runtime/test_custom_provider_lifecycle.py
index 96b78352..647a56a8 100644
--- a/tests/runtime/test_custom_provider_lifecycle.py
+++ b/tests/runtime/test_custom_provider_lifecycle.py
@@ -6,6 +6,8 @@ from free_claude_code.application.model_catalog import read_model_catalog
from free_claude_code.application.model_metadata import ProviderModelInfo
from free_claude_code.config.custom_providers import CustomProviderDefinition
from free_claude_code.config.settings import Settings
+from free_claude_code.providers.admission_policy import ProviderAdmissionLimits
+from free_claude_code.providers.admission_registry import ProviderAdmissionRegistry
from free_claude_code.providers.custom import CustomProvider
from free_claude_code.providers.runtime import ProviderRuntime
from free_claude_code.runtime.provider_manager import ProviderRuntimeManager
@@ -32,7 +34,12 @@ async def commit():
async def test_runtime_constructs_one_custom_owner_lazily():
- runtime = ProviderRuntime(settings(model_ids=["manual"]))
+ runtime = ProviderRuntime(
+ settings(model_ids=["manual"]),
+ ProviderAdmissionRegistry(
+ ProviderAdmissionLimits.from_settings(settings(model_ids=["manual"]))
+ ),
+ )
try:
assert not runtime.is_cached(ID)
one, two = await asyncio.gather(
@@ -70,8 +77,8 @@ async def test_changed_endpoint_cannot_inherit_catalog_but_rename_can():
await release.wait()
return frozenset()
- def blocked(settings):
- runtime = factory(settings)
+ def blocked(settings, admission_registry):
+ runtime = factory(settings, admission_registry)
assert isinstance(runtime, FakeRuntime)
runtime.provider.list_model_infos.side_effect = wait
return runtime
diff --git a/tests/runtime/test_integration_startup.py b/tests/runtime/test_integration_startup.py
index d0ce318f..9efe7cbe 100644
--- a/tests/runtime/test_integration_startup.py
+++ b/tests/runtime/test_integration_startup.py
@@ -50,13 +50,13 @@ def runtime():
return_value=frozenset({ProviderModelInfo("one")})
)
- async def construct(_id, _settings):
+ async def construct(_id, _settings, _admission_registry):
return provider
manager = ProviderRuntimeManager(
settings,
- runtime_factory=lambda snapshot: ProviderRuntime(
- snapshot, provider_constructor=construct
+ runtime_factory=lambda snapshot, admission_registry: ProviderRuntime(
+ snapshot, admission_registry, provider_constructor=construct
),
)
return ApplicationRuntime(
diff --git a/tests/runtime/test_progressive_startup.py b/tests/runtime/test_progressive_startup.py
index c166ca0c..ede123bb 100644
--- a/tests/runtime/test_progressive_startup.py
+++ b/tests/runtime/test_progressive_startup.py
@@ -47,13 +47,15 @@ async def test_http_and_admin_serve_while_selected_catalog_is_held():
provider.list_model_infos = AsyncMock(side_effect=list_models)
- async def construct(_provider_id, _settings):
+ async def construct(_provider_id, _settings, _admission_registry):
return provider
manager = ProviderRuntimeManager(
_settings(),
- runtime_factory=lambda settings: ProviderRuntime(
- settings, provider_constructor=construct
+ runtime_factory=lambda settings, admission_registry: ProviderRuntime(
+ settings,
+ admission_registry,
+ provider_constructor=construct,
),
)
runtime = ApplicationRuntime(
@@ -94,13 +96,15 @@ async def test_independent_provider_and_cancelled_waiter_share_initialization():
return_value=frozenset({ProviderModelInfo("one")})
)
- async def construct(provider_id, _settings):
+ async def construct(provider_id, _settings, _admission_registry):
return slow if provider_id == "groq" else fast
manager = ProviderRuntimeManager(
_settings(),
- runtime_factory=lambda settings: ProviderRuntime(
- settings, provider_constructor=construct
+ runtime_factory=lambda settings, admission_registry: ProviderRuntime(
+ settings,
+ admission_registry,
+ provider_constructor=construct,
),
)
manager.start_model_list_refresh()
@@ -153,7 +157,7 @@ async def test_failed_discovery_construction_recovers_in_same_generation(refresh
)
attempts = 0
- async def construct(_id, _settings):
+ async def construct(_id, _settings, _admission_registry):
nonlocal attempts
attempts += 1
if attempts == 1:
@@ -162,8 +166,10 @@ async def test_failed_discovery_construction_recovers_in_same_generation(refresh
manager = ProviderRuntimeManager(
_settings(),
- runtime_factory=lambda settings: ProviderRuntime(
- settings, provider_constructor=construct
+ runtime_factory=lambda settings, admission_registry: ProviderRuntime(
+ settings,
+ admission_registry,
+ provider_constructor=construct,
),
)
generation = manager.current_generation_id
@@ -193,7 +199,7 @@ async def test_recovery_wait_uses_remaining_budget_without_orphaning_waiter(
attempts = 0
construction = None
- async def construct(_id, _settings):
+ async def construct(_id, _settings, _admission_registry):
nonlocal attempts, construction
attempts += 1
if attempts == 1:
@@ -205,8 +211,10 @@ async def test_recovery_wait_uses_remaining_budget_without_orphaning_waiter(
manager = ProviderRuntimeManager(
_settings(),
- runtime_factory=lambda settings: ProviderRuntime(
- settings, provider_constructor=construct
+ runtime_factory=lambda settings, admission_registry: ProviderRuntime(
+ settings,
+ admission_registry,
+ provider_constructor=construct,
),
)
wait = InitializationWait(1 if cancel else 0.01)
@@ -269,7 +277,7 @@ async def test_late_retired_discovery_cannot_replace_current_metadata():
entered, release = asyncio.Event(), asyncio.Event()
providers = []
- def factory(settings):
+ def factory(settings, admission_registry):
provider = MagicMock(spec=BaseProvider)
providers.append(provider)
ordinal = len(providers)
@@ -282,10 +290,14 @@ async def test_late_retired_discovery_cannot_replace_current_metadata():
provider.list_model_infos = AsyncMock(side_effect=models)
- async def construct(_provider_id, _settings):
+ async def construct(_provider_id, _settings, _admission_registry):
return provider
- return ProviderRuntime(settings, provider_constructor=construct)
+ return ProviderRuntime(
+ settings,
+ admission_registry,
+ provider_constructor=construct,
+ )
manager = ProviderRuntimeManager(_settings(), runtime_factory=factory)
old = await manager.acquire()
@@ -326,13 +338,15 @@ async def test_complete_catalog_publication_preserves_prior_file_until_all_attem
provider.list_model_infos = AsyncMock(side_effect=models)
- async def construct(_provider_id, _settings):
+ async def construct(_provider_id, _settings, _admission_registry):
return provider
manager = ProviderRuntimeManager(
_settings(),
- runtime_factory=lambda settings: ProviderRuntime(
- settings, provider_constructor=construct
+ runtime_factory=lambda settings, admission_registry: ProviderRuntime(
+ settings,
+ admission_registry,
+ provider_constructor=construct,
),
model_catalog_publisher=CodexModelCatalogPublisher(path),
)
@@ -360,7 +374,7 @@ async def test_catalog_wait_follows_replacement(
entered, release = asyncio.Event(), asyncio.Event()
count = 0
- def runtime(settings):
+ def runtime(settings, admission_registry):
nonlocal count
count += 1
old = count == 1
@@ -374,10 +388,14 @@ async def test_catalog_wait_follows_replacement(
provider.list_model_infos = AsyncMock(side_effect=models)
- async def construct(_id, _settings):
+ async def construct(_id, _settings, _admission_registry):
return provider
- return ProviderRuntime(settings, provider_constructor=construct)
+ return ProviderRuntime(
+ settings,
+ admission_registry,
+ provider_constructor=construct,
+ )
path = tmp_path / "catalog.json"
manager = ProviderRuntimeManager(
@@ -442,13 +460,15 @@ async def test_catalog_waiter_exit_preserves_shared_discovery(
provider.list_model_infos = AsyncMock(side_effect=models)
- async def construct(_id, _settings):
+ async def construct(_id, _settings, _admission_registry):
return provider
manager = ProviderRuntimeManager(
_settings(),
- runtime_factory=lambda settings: ProviderRuntime(
- settings, provider_constructor=construct
+ runtime_factory=lambda settings, admission_registry: ProviderRuntime(
+ settings,
+ admission_registry,
+ provider_constructor=construct,
),
)
wait = InitializationWait(1 if cancel else 0.01)
diff --git a/tests/runtime/test_provider_manager.py b/tests/runtime/test_provider_manager.py
index 616b977d..5270b67a 100644
--- a/tests/runtime/test_provider_manager.py
+++ b/tests/runtime/test_provider_manager.py
@@ -45,7 +45,7 @@ class RuntimeFactory:
self.runtimes: list[FakeRuntime] = []
self.error: Exception | None = None
- def __call__(self, settings: Settings) -> ProviderRuntime:
+ def __call__(self, settings: Settings, admission_registry) -> ProviderRuntime:
if self.error is not None:
raise self.error
runtime = FakeRuntime(settings)
@@ -302,7 +302,7 @@ async def test_replacement_keeps_leased_generation_open_until_final_release() ->
@pytest.mark.asyncio
-async def test_hot_replacement_owns_admission_per_provider_generation() -> None:
+async def test_hot_replacement_shares_admission_and_retires_clients() -> None:
first_settings = _settings("nvidia_nim/one")
second_settings = _settings("nvidia_nim/two")
clients: list[MagicMock] = []
@@ -331,7 +331,7 @@ async def test_hot_replacement_owns_admission_per_provider_generation() -> None:
assert isinstance(old_provider, NvidiaNimProvider)
assert isinstance(new_provider, NvidiaNimProvider)
assert new_provider is not old_provider
- assert new_provider._admission is not old_provider._admission
+ assert new_provider._admission is old_provider._admission
assert (await old_lease.resolve_provider("nvidia_nim")) is old_provider
clients[0].close.assert_not_awaited()
diff --git a/tests/runtime/test_retired_chat.py b/tests/runtime/test_retired_chat.py
index 80a9dfa2..4f1151c4 100644
--- a/tests/runtime/test_retired_chat.py
+++ b/tests/runtime/test_retired_chat.py
@@ -40,7 +40,9 @@ def _runtime():
store.initialize()
manager = ProviderRuntimeManager(
store.read().settings,
- runtime_factory=lambda snapshot: ProviderRuntime(snapshot, {}),
+ runtime_factory=lambda snapshot, admission_registry: ProviderRuntime(
+ snapshot, admission_registry, {}
+ ),
)
return ApplicationRuntime(
manager, configuration=ConfigurationService(store), transcriber=None
diff --git a/tests/runtime/test_shared_provider_admission.py b/tests/runtime/test_shared_provider_admission.py
new file mode 100644
index 00000000..8c623159
--- /dev/null
+++ b/tests/runtime/test_shared_provider_admission.py
@@ -0,0 +1,266 @@
+"""Admission protections survive replacement of their provider clients."""
+
+import asyncio
+from typing import cast
+from unittest.mock import AsyncMock, patch
+
+import httpx
+import pytest
+
+from free_claude_code.config.settings import Settings
+from free_claude_code.providers.admission import ProviderOperationKind
+from free_claude_code.providers.admission_policy import ProviderAdmissionLimits
+from free_claude_code.providers.admission_registry import ProviderAdmissionRegistry
+from free_claude_code.providers.nvidia_nim import NvidiaNimProvider
+from free_claude_code.runtime.provider_manager import ProviderRuntimeManager
+from tests.runtime.test_provider_manager import RuntimeFactory
+
+
[email protected]("protection", ["concurrency", "rate", "recovery"])
+async def test_replacement_cannot_bypass_existing_protection(protection: str) -> None:
+ settings = Settings(
+ nvidia_nim_api_key="test-key",
+ provider_rate_limit=1 if protection == "rate" else 1000,
+ provider_rate_window=60,
+ provider_max_concurrency=1,
+ )
+ manager = ProviderRuntimeManager(settings)
+ old_lease = await manager.acquire()
+ pending = None
+ old_attempt = None
+ try:
+ old = cast(
+ NvidiaNimProvider,
+ await manager._current.runtime.resolve_provider("nvidia_nim"),
+ )
+ old_attempt = await old._admission.start_execution().open_attempt(
+ ProviderOperationKind.GENERATION
+ )
+ if protection == "recovery":
+ response = httpx.Response(
+ 503,
+ headers={"retry-after": "60"},
+ request=httpx.Request("POST", "https://provider.test"),
+ )
+ await old_attempt.fail(
+ httpx.HTTPStatusError(
+ "unavailable", request=response.request, response=response
+ )
+ )
+ else:
+ await old_attempt.accept()
+ if protection != "concurrency":
+ await old_attempt.aclose()
+
+ with patch.object(
+ NvidiaNimProvider, "list_model_infos", AsyncMock(return_value=frozenset())
+ ):
+ await manager.replace(
+ settings.model_copy(update={"log_raw_sse_events": True}),
+ commit=AsyncMock(),
+ )
+ new = cast(
+ NvidiaNimProvider,
+ await manager._current.runtime.resolve_provider("nvidia_nim"),
+ )
+ pending = asyncio.create_task(
+ new._admission.start_execution().open_attempt(
+ ProviderOperationKind.GENERATION
+ )
+ )
+ await asyncio.sleep(0)
+ assert not pending.done(), f"replacement bypassed {protection}"
+ finally:
+ if pending is not None:
+ pending.cancel()
+ outcomes = await asyncio.gather(pending, return_exceptions=True)
+ if not isinstance(outcomes[0], BaseException):
+ await outcomes[0].aclose()
+ if old_attempt is not None:
+ await old_attempt.aclose()
+ await old_lease.release()
+ await manager.close()
+
+
[email protected]("outcome", ["success", "failure", "cancel"])
+async def test_limits_publish_only_with_successful_settings(outcome):
+ settings = Settings(provider_rate_limit=1000, provider_max_concurrency=1)
+ manager = ProviderRuntimeManager(settings, runtime_factory=RuntimeFactory())
+ controller = manager._admission_registry.get("nvidia_nim")
+ first = await controller.start_execution().open_attempt(
+ ProviderOperationKind.GENERATION
+ )
+ pending = asyncio.create_task(
+ controller.start_execution().open_attempt(ProviderOperationKind.GENERATION)
+ )
+ entered, release = asyncio.Event(), asyncio.Event()
+
+ async def commit():
+ entered.set()
+ await release.wait()
+ if outcome == "failure":
+ raise OSError("cannot save")
+
+ replacement = asyncio.create_task(
+ manager.replace(
+ settings.model_copy(update={"provider_max_concurrency": 2}), commit=commit
+ )
+ )
+ await entered.wait()
+ assert not pending.done()
+ if outcome == "cancel":
+ replacement.cancel()
+ else:
+ release.set()
+ results = await asyncio.gather(replacement, return_exceptions=True)
+ if outcome == "success":
+ assert results == [2]
+ await (await asyncio.wait_for(pending, 1)).aclose()
+ else:
+ assert isinstance(results[0], BaseException)
+ await asyncio.sleep(0)
+ assert not pending.done()
+ pending.cancel()
+ await asyncio.gather(pending, return_exceptions=True)
+ await first.aclose()
+ await manager.close()
+
+
+async def test_blocked_publication_keeps_old_limits_until_generation_switch():
+ settings = Settings(provider_rate_limit=1000, provider_max_concurrency=1)
+ manager = ProviderRuntimeManager(settings, runtime_factory=RuntimeFactory())
+ controller = manager._admission_registry.get("nvidia_nim")
+ first = await controller.start_execution().open_attempt(
+ ProviderOperationKind.GENERATION
+ )
+ pending = asyncio.create_task(
+ controller.start_execution().open_attempt(ProviderOperationKind.GENERATION)
+ )
+ committed = asyncio.Event()
+
+ async def commit():
+ committed.set()
+
+ await manager._publication_lock.acquire()
+ replacement = asyncio.create_task(
+ manager.replace(
+ settings.model_copy(update={"provider_max_concurrency": 2}), commit=commit
+ )
+ )
+ await committed.wait()
+ assert not pending.done()
+ assert manager.current_generation_id == 1
+ manager._publication_lock.release()
+ await replacement
+ await (await asyncio.wait_for(pending, 1)).aclose()
+ await first.aclose()
+ await manager.close()
+
+
+async def test_old_lazy_construction_uses_current_limits_and_survives_retirement():
+ settings = Settings(
+ nvidia_nim_api_key="test", provider_rate_limit=1000, provider_max_concurrency=1
+ )
+ manager = ProviderRuntimeManager(settings)
+ old_lease = await manager.acquire()
+ old_runtime = manager._current.runtime
+ with patch.object(
+ NvidiaNimProvider, "list_model_infos", AsyncMock(return_value=frozenset())
+ ):
+ await manager.replace(
+ settings.model_copy(update={"provider_max_concurrency": 2}),
+ commit=AsyncMock(),
+ )
+ old = cast(NvidiaNimProvider, await old_runtime.resolve_provider("nvidia_nim"))
+ new = cast(
+ NvidiaNimProvider,
+ await manager._current.runtime.resolve_provider("nvidia_nim"),
+ )
+ first = await old._admission.start_execution().open_attempt(
+ ProviderOperationKind.GENERATION
+ )
+ second = await asyncio.wait_for(
+ new._admission.start_execution().open_attempt(
+ ProviderOperationKind.GENERATION
+ ),
+ 1,
+ )
+ blocked = asyncio.create_task(
+ new._admission.start_execution().open_attempt(
+ ProviderOperationKind.GENERATION
+ )
+ )
+ await asyncio.sleep(0)
+ assert not blocked.done()
+ blocked.cancel()
+ await asyncio.gather(blocked, return_exceptions=True)
+ await first.aclose()
+ await second.aclose()
+ await old_lease.release()
+ await asyncio.wait_for(
+ new._admission.start_execution().run_call(
+ AsyncMock(return_value="ok"),
+ operation_kind=ProviderOperationKind.GENERATION,
+ ),
+ 1,
+ )
+ await manager.close()
+
+
+async def test_separate_provider_ids_and_managers_have_independent_capacity():
+ limits = ProviderAdmissionLimits(1, 60, 1)
+ registry = ProviderAdmissionRegistry(limits)
+ other = ProviderAdmissionRegistry(limits)
+ attempts = []
+ for owner, provider in [(registry, "one"), (registry, "two"), (other, "one")]:
+ attempts.append(
+ await asyncio.wait_for(
+ owner.get(provider)
+ .start_execution()
+ .open_attempt(ProviderOperationKind.GENERATION),
+ 1,
+ )
+ )
+ for attempt in attempts:
+ await attempt.aclose()
+ registry.close()
+ other.close()
+
+
+async def test_custom_registry_entry_lives_until_retired_generation_cleanup_succeeds():
+ from free_claude_code.config.custom_providers import CustomProviderDefinition
+
+ custom_id = "custom_12345678123412341234123456789abc"
+ settings = Settings(
+ custom_providers=(
+ CustomProviderDefinition(
+ provider_id=custom_id,
+ display_name="Example",
+ base_url="https://example.test/v1",
+ api_format="openai_chat",
+ ),
+ )
+ )
+ factory = RuntimeFactory()
+ manager = ProviderRuntimeManager(settings, runtime_factory=factory)
+ lease = await manager.acquire()
+ controller = manager._admission_registry.get(custom_id)
+ await manager.replace(
+ settings.model_copy(update={"custom_providers": ()}), commit=AsyncMock()
+ )
+ assert manager._admission_registry.get(custom_id) is controller
+ factory.runtimes[0].cleanup_error = OSError("not closed yet")
+ await lease.release()
+ assert manager._admission_registry.get(custom_id) is controller
+ factory.runtimes[0].cleanup_error = None
+ await manager._close_generation(manager._retired[1], forced=False)
+ assert custom_id not in manager._admission_registry._controllers
+ await manager.close()
+ with pytest.raises(RuntimeError, match="closed"):
+ manager._admission_registry.get("nvidia_nim")
diff --git a/tests/runtime/test_vscode_chat_sync.py b/tests/runtime/test_vscode_chat_sync.py
index 89e95f74..a6847f4b 100644
--- a/tests/runtime/test_vscode_chat_sync.py
+++ b/tests/runtime/test_vscode_chat_sync.py
@@ -35,7 +35,11 @@ async def runtime(request):
manager = ProviderRuntimeManager(
settings,
- runtime_factory=lambda s: ProviderRuntime(s, provider_constructor=construct),
+ runtime_factory=lambda s, admission_registry: ProviderRuntime(
+ s,
+ admission_registry,
+ provider_constructor=construct,
+ ),
)
app = ApplicationRuntime(
manager,