fix linter

api/enums/quota_type.py

@@ -1,56 +1,17 @@
-import logging
-from dataclasses import dataclass
 from enum import StrEnum, auto
 
-logger = logging.getLogger(__name__)
-
-
-@dataclass
-class QuotaCharge:
-    """
-    Result of a quota consumption operation.
-
-    Attributes:
-        success: Whether the quota charge succeeded
-        charge_id: UUID for refund, or None if failed/disabled
-    """
-
-    success: bool
-    charge_id: str | None
-    _quota_type: "QuotaType"
-
-    def refund(self) -> None:
-        """
-        Refund this quota charge.
-
-        Safe to call even if charge failed or was disabled.
-        This method guarantees no exceptions will be raised.
-        """
-        if self.charge_id:
-            self._quota_type.refund(self.charge_id)
-            logger.info("Refunded quota for %s with charge_id: %s", self._quota_type.value, self.charge_id)
-
 
 class QuotaType(StrEnum):
     """
     Supported quota types for tenant feature usage.
-
-    Add additional types here whenever new billable features become available.
     """
 
-    # Trigger execution quota
     TRIGGER = auto()
-
-    # Workflow execution quota
     WORKFLOW = auto()
-
     UNLIMITED = auto()
 
     @property
     def billing_key(self) -> str:
-        """
-        Get the billing key for the feature.
-        """
         match self:
             case QuotaType.TRIGGER:
                 return "trigger_event"
@@ -58,152 +19,3 @@ class QuotaType(StrEnum):
                 return "api_rate_limit"
             case _:
                 raise ValueError(f"Invalid quota type: {self}")
-
-    def consume(self, tenant_id: str, amount: int = 1) -> QuotaCharge:
-        """
-        Consume quota for the feature.
-
-        Args:
-            tenant_id: The tenant identifier
-            amount: Amount to consume (default: 1)
-
-        Returns:
-            QuotaCharge with success status and charge_id for refund
-
-        Raises:
-            QuotaExceededError: When quota is insufficient
-        """
-        from configs import dify_config
-        from services.billing_service import BillingService
-        from services.errors.app import QuotaExceededError
-
-        if not dify_config.BILLING_ENABLED:
-            logger.debug("Billing disabled, allowing request for %s", tenant_id)
-            return QuotaCharge(success=True, charge_id=None, _quota_type=self)
-
-        logger.info("Consuming %d %s quota for tenant %s", amount, self.value, tenant_id)
-
-        if amount <= 0:
-            raise ValueError("Amount to consume must be greater than 0")
-
-        try:
-            response = BillingService.update_tenant_feature_plan_usage(tenant_id, self.billing_key, delta=amount)
-
-            if response.get("result") != "success":
-                logger.warning(
-                    "Failed to consume quota for %s, feature %s details: %s",
-                    tenant_id,
-                    self.value,
-                    response.get("detail"),
-                )
-                raise QuotaExceededError(feature=self.value, tenant_id=tenant_id, required=amount)
-
-            charge_id = response.get("history_id")
-            logger.debug(
-                "Successfully consumed %d %s quota for tenant %s, charge_id: %s",
-                amount,
-                self.value,
-                tenant_id,
-                charge_id,
-            )
-            return QuotaCharge(success=True, charge_id=charge_id, _quota_type=self)
-
-        except QuotaExceededError:
-            raise
-        except Exception:
-            # fail-safe: allow request on billing errors
-            logger.exception("Failed to consume quota for %s, feature %s", tenant_id, self.value)
-            return unlimited()
-
-    def check(self, tenant_id: str, amount: int = 1) -> bool:
-        """
-        Check if tenant has sufficient quota without consuming.
-
-        Args:
-            tenant_id: The tenant identifier
-            amount: Amount to check (default: 1)
-
-        Returns:
-            True if quota is sufficient, False otherwise
-        """
-        from configs import dify_config
-
-        if not dify_config.BILLING_ENABLED:
-            return True
-
-        if amount <= 0:
-            raise ValueError("Amount to check must be greater than 0")
-
-        try:
-            remaining = self.get_remaining(tenant_id)
-            return remaining >= amount if remaining != -1 else True
-        except Exception:
-            logger.exception("Failed to check quota for %s, feature %s", tenant_id, self.value)
-            # fail-safe: allow request on billing errors
-            return True
-
-    def refund(self, charge_id: str) -> None:
-        """
-        Refund quota using charge_id from consume().
-
-        This method guarantees no exceptions will be raised.
-        All errors are logged but silently handled.
-
-        Args:
-            charge_id: The UUID returned from consume()
-        """
-        try:
-            from configs import dify_config
-            from services.billing_service import BillingService
-
-            if not dify_config.BILLING_ENABLED:
-                return
-
-            if not charge_id:
-                logger.warning("Cannot refund: charge_id is empty")
-                return
-
-            logger.info("Refunding %s quota with charge_id: %s", self.value, charge_id)
-
-            response = BillingService.refund_tenant_feature_plan_usage(charge_id)
-            if response.get("result") == "success":
-                logger.debug("Successfully refunded %s quota, charge_id: %s", self.value, charge_id)
-            else:
-                logger.warning("Refund failed for charge_id: %s", charge_id)
-
-        except Exception:
-            # Catch ALL exceptions - refund must never fail
-            logger.exception("Failed to refund quota for charge_id: %s", charge_id)
-            # Don't raise - refund is best-effort and must be silent
-
-    def get_remaining(self, tenant_id: str) -> int:
-        """
-        Get remaining quota for the tenant.
-
-        Args:
-            tenant_id: The tenant identifier
-
-        Returns:
-            Remaining quota amount
-        """
-        from services.billing_service import BillingService
-
-        try:
-            usage_info = BillingService.get_tenant_feature_plan_usage(tenant_id, self.billing_key)
-            # Assuming the API returns a dict with 'remaining' or 'limit' and 'used'
-            if isinstance(usage_info, dict):
-                return usage_info.get("remaining", 0)
-            # If it returns a simple number, treat it as remaining
-            return int(usage_info) if usage_info else 0
-        except Exception:
-            logger.exception("Failed to get remaining quota for %s, feature %s", tenant_id, self.value)
-            return -1
-
-
-def unlimited() -> QuotaCharge:
-    """
-    Return a quota charge for unlimited quota.
-
-    This is useful for features that are not subject to quota limits, such as the UNLIMITED quota type.
-    """
-    return QuotaCharge(success=True, charge_id=None, _quota_type=QuotaType.UNLIMITED)

api/services/app_generate_service.py

@@ -18,12 +18,13 @@ from core.app.features.rate_limiting import RateLimit
 from core.app.features.rate_limiting.rate_limit import rate_limit_context
 from core.app.layers.pause_state_persist_layer import PauseStateLayerConfig
 from core.db import session_factory
-from enums.quota_type import QuotaType, unlimited
+from enums.quota_type import QuotaType
 from extensions.otel import AppGenerateHandler, trace_span
 from models.model import Account, App, AppMode, EndUser
 from models.workflow import Workflow, WorkflowRun
 from services.errors.app import QuotaExceededError, WorkflowIdFormatError, WorkflowNotFoundError
 from services.errors.llm import InvokeRateLimitError
+from services.quota_service import QuotaService, unlimited
 from services.workflow_service import WorkflowService
 from tasks.app_generate.workflow_execute_task import AppExecutionParams, workflow_based_app_execution_task
 
@@ -106,7 +107,7 @@ class AppGenerateService:
         quota_charge = unlimited()
         if dify_config.BILLING_ENABLED:
             try:
-                quota_charge = QuotaType.WORKFLOW.consume(app_model.tenant_id)
+                quota_charge = QuotaService.reserve(QuotaType.WORKFLOW, app_model.tenant_id)
             except QuotaExceededError:
                 raise InvokeRateLimitError(f"Workflow execution quota limit reached for tenant {app_model.tenant_id}")
 
@@ -116,6 +117,7 @@ class AppGenerateService:
         request_id = RateLimit.gen_request_key()
         try:
             request_id = rate_limit.enter(request_id)
+            quota_charge.commit()
             if app_model.mode == AppMode.COMPLETION:
                 return rate_limit.generate(
                     CompletionAppGenerator.convert_to_event_stream(

api/services/async_workflow_service.py

@@ -22,6 +22,7 @@ from models.trigger import WorkflowTriggerLog, WorkflowTriggerLogDict
 from models.workflow import Workflow
 from repositories.sqlalchemy_workflow_trigger_log_repository import SQLAlchemyWorkflowTriggerLogRepository
 from services.errors.app import QuotaExceededError, WorkflowNotFoundError, WorkflowQuotaLimitError
+from services.quota_service import QuotaService, unlimited
 from services.workflow.entities import AsyncTriggerResponse, TriggerData, WorkflowTaskData
 from services.workflow.queue_dispatcher import QueueDispatcherManager, QueuePriority
 from services.workflow_service import WorkflowService
@@ -131,9 +132,10 @@ class AsyncWorkflowService:
         trigger_log = trigger_log_repo.create(trigger_log)
         session.commit()
 
-        # 7. Check and consume quota
+        # 7. Reserve quota (commit after successful dispatch)
+        quota_charge = unlimited()
         try:
-            QuotaType.WORKFLOW.consume(trigger_data.tenant_id)
+            quota_charge = QuotaService.reserve(QuotaType.WORKFLOW, trigger_data.tenant_id)
         except QuotaExceededError as e:
             # Update trigger log status
             trigger_log.status = WorkflowTriggerStatus.RATE_LIMITED
@@ -153,13 +155,18 @@ class AsyncWorkflowService:
         # 9. Dispatch to appropriate queue
         task_data_dict = task_data.model_dump(mode="json")
 
-        task: AsyncResult[Any] | None = None
-        if queue_name == QueuePriority.PROFESSIONAL:
-            task = execute_workflow_professional.delay(task_data_dict)
-        elif queue_name == QueuePriority.TEAM:
-            task = execute_workflow_team.delay(task_data_dict)
-        else:  # SANDBOX
-            task = execute_workflow_sandbox.delay(task_data_dict)
+        try:
+            task: AsyncResult[Any] | None = None
+            if queue_name == QueuePriority.PROFESSIONAL:
+                task = execute_workflow_professional.delay(task_data_dict)
+            elif queue_name == QueuePriority.TEAM:
+                task = execute_workflow_team.delay(task_data_dict)
+            else:  # SANDBOX
+                task = execute_workflow_sandbox.delay(task_data_dict)
+            quota_charge.commit()
+        except Exception:
+            quota_charge.refund()
+            raise
 
         # 10. Update trigger log with task info
         trigger_log.status = WorkflowTriggerStatus.QUEUED

api/services/billing_service.py

@@ -26,6 +26,29 @@ class SubscriptionPlan(TypedDict):
     expiration_date: int
 
 
+class QuotaReserveResult(TypedDict):
+    reservation_id: str
+    available: int
+    reserved: int
+
+
+class QuotaCommitResult(TypedDict):
+    available: int
+    reserved: int
+    refunded: int
+
+
+class QuotaReleaseResult(TypedDict):
+    available: int
+    reserved: int
+    released: int
+
+
+_quota_reserve_adapter = TypeAdapter(QuotaReserveResult)
+_quota_commit_adapter = TypeAdapter(QuotaCommitResult)
+_quota_release_adapter = TypeAdapter(QuotaReleaseResult)
+
+
 class BillingService:
     base_url = os.environ.get("BILLING_API_URL", "BILLING_API_URL")
     secret_key = os.environ.get("BILLING_API_SECRET_KEY", "BILLING_API_SECRET_KEY")
@@ -46,11 +69,61 @@ class BillingService:
 
     @classmethod
     def get_tenant_feature_plan_usage_info(cls, tenant_id: str):
+        """Deprecated: Use get_quota_info instead."""
         params = {"tenant_id": tenant_id}
-
         usage_info = cls._send_request("GET", "/tenant-feature-usage/info", params=params)
         return usage_info
 
+    @classmethod
+    def get_quota_info(cls, tenant_id: str):
+        params = {"tenant_id": tenant_id}
+        return cls._send_request("GET", "/quota/info", params=params)
+
+    @classmethod
+    def quota_reserve(
+        cls, tenant_id: str, feature_key: str, request_id: str, amount: int = 1, meta: dict | None = None
+    ) -> QuotaReserveResult:
+        """Reserve quota before task execution."""
+        payload: dict = {
+            "tenant_id": tenant_id,
+            "feature_key": feature_key,
+            "request_id": request_id,
+            "amount": amount,
+        }
+        if meta:
+            payload["meta"] = meta
+        return _quota_reserve_adapter.validate_python(cls._send_request("POST", "/quota/reserve", json=payload))
+
+    @classmethod
+    def quota_commit(
+        cls, tenant_id: str, feature_key: str, reservation_id: str, actual_amount: int, meta: dict | None = None
+    ) -> QuotaCommitResult:
+        """Commit a reservation with actual consumption."""
+        payload: dict = {
+            "tenant_id": tenant_id,
+            "feature_key": feature_key,
+            "reservation_id": reservation_id,
+            "actual_amount": actual_amount,
+        }
+        if meta:
+            payload["meta"] = meta
+        return _quota_commit_adapter.validate_python(cls._send_request("POST", "/quota/commit", json=payload))
+
+    @classmethod
+    def quota_release(cls, tenant_id: str, feature_key: str, reservation_id: str) -> QuotaReleaseResult:
+        """Release a reservation (cancel, return frozen quota)."""
+        return _quota_release_adapter.validate_python(
+            cls._send_request(
+                "POST",
+                "/quota/release",
+                json={
+                    "tenant_id": tenant_id,
+                    "feature_key": feature_key,
+                    "reservation_id": reservation_id,
+                },
+            )
+        )
+
     @classmethod
     def get_knowledge_rate_limit(cls, tenant_id: str):
         params = {"tenant_id": tenant_id}

api/services/quota_service.py

@@ -0,0 +1,233 @@
+from __future__ import annotations
+
+import logging
+import uuid
+from dataclasses import dataclass, field
+from typing import TYPE_CHECKING
+
+from configs import dify_config
+
+if TYPE_CHECKING:
+    from enums.quota_type import QuotaType
+
+logger = logging.getLogger(__name__)
+
+
+@dataclass
+class QuotaCharge:
+    """
+    Result of a quota reservation (Reserve phase).
+
+    Lifecycle:
+        charge = QuotaService.consume(QuotaType.TRIGGER, tenant_id)
+        try:
+            do_work()
+            charge.commit()   # Confirm consumption
+        except:
+            charge.refund()   # Release frozen quota
+
+    If neither commit() nor refund() is called, the billing system's
+    cleanup CronJob will auto-release the reservation within ~75 seconds.
+    """
+
+    success: bool
+    charge_id: str | None  # reservation_id
+    _quota_type: QuotaType
+    _tenant_id: str | None = None
+    _feature_key: str | None = None
+    _amount: int = 0
+    _committed: bool = field(default=False, repr=False)
+
+    def commit(self, actual_amount: int | None = None) -> None:
+        """
+        Confirm the consumption with actual amount.
+
+        Args:
+            actual_amount: Actual amount consumed. Defaults to the reserved amount.
+                           If less than reserved, the difference is refunded automatically.
+        """
+        if self._committed or not self.charge_id or not self._tenant_id or not self._feature_key:
+            return
+
+        try:
+            from services.billing_service import BillingService
+
+            amount = actual_amount if actual_amount is not None else self._amount
+            BillingService.quota_commit(
+                tenant_id=self._tenant_id,
+                feature_key=self._feature_key,
+                reservation_id=self.charge_id,
+                actual_amount=amount,
+            )
+            self._committed = True
+            logger.debug(
+                "Committed %s quota for tenant %s, reservation_id: %s, amount: %d",
+                self._quota_type,
+                self._tenant_id,
+                self.charge_id,
+                amount,
+            )
+        except Exception:
+            logger.exception("Failed to commit quota, reservation_id: %s", self.charge_id)
+
+    def refund(self) -> None:
+        """
+        Release the reserved quota (cancel the charge).
+
+        Safe to call even if:
+        - charge failed or was disabled (charge_id is None)
+        - already committed (Release after Commit is a no-op)
+        - already refunded (idempotent)
+
+        This method guarantees no exceptions will be raised.
+        """
+        if not self.charge_id or not self._tenant_id or not self._feature_key:
+            return
+
+        QuotaService.release(self._quota_type, self.charge_id, self._tenant_id, self._feature_key)
+
+
+def unlimited() -> QuotaCharge:
+    from enums.quota_type import QuotaType
+
+    return QuotaCharge(success=True, charge_id=None, _quota_type=QuotaType.UNLIMITED)
+
+
+class QuotaService:
+    """Orchestrates quota reserve / commit / release lifecycle via BillingService."""
+
+    @staticmethod
+    def consume(quota_type: QuotaType, tenant_id: str, amount: int = 1) -> QuotaCharge:
+        """
+        Reserve + immediate Commit (one-shot mode).
+
+        The returned QuotaCharge supports .refund() which calls Release.
+        For two-phase usage (e.g. streaming), use reserve() directly.
+        """
+        charge = QuotaService.reserve(quota_type, tenant_id, amount)
+        if charge.success and charge.charge_id:
+            charge.commit()
+        return charge
+
+    @staticmethod
+    def reserve(quota_type: QuotaType, tenant_id: str, amount: int = 1) -> QuotaCharge:
+        """
+        Reserve quota before task execution (Reserve phase only).
+
+        The caller MUST call charge.commit() after the task succeeds,
+        or charge.refund() if the task fails.
+
+        Raises:
+            QuotaExceededError: When quota is insufficient
+        """
+        from services.billing_service import BillingService
+        from services.errors.app import QuotaExceededError
+
+        if not dify_config.BILLING_ENABLED:
+            logger.debug("Billing disabled, allowing request for %s", tenant_id)
+            return QuotaCharge(success=True, charge_id=None, _quota_type=quota_type)
+
+        logger.info("Reserving %d %s quota for tenant %s", amount, quota_type.value, tenant_id)
+
+        if amount <= 0:
+            raise ValueError("Amount to reserve must be greater than 0")
+
+        request_id = str(uuid.uuid4())
+        feature_key = quota_type.billing_key
+
+        try:
+            reserve_resp = BillingService.quota_reserve(
+                tenant_id=tenant_id,
+                feature_key=feature_key,
+                request_id=request_id,
+                amount=amount,
+            )
+
+            reservation_id = reserve_resp.get("reservation_id")
+            if not reservation_id:
+                logger.warning(
+                    "Reserve returned no reservation_id for %s, feature %s, response: %s",
+                    tenant_id,
+                    quota_type.value,
+                    reserve_resp,
+                )
+                raise QuotaExceededError(feature=quota_type.value, tenant_id=tenant_id, required=amount)
+
+            logger.debug(
+                "Reserved %d %s quota for tenant %s, reservation_id: %s",
+                amount,
+                quota_type.value,
+                tenant_id,
+                reservation_id,
+            )
+            return QuotaCharge(
+                success=True,
+                charge_id=reservation_id,
+                _quota_type=quota_type,
+                _tenant_id=tenant_id,
+                _feature_key=feature_key,
+                _amount=amount,
+            )
+
+        except QuotaExceededError:
+            raise
+        except ValueError:
+            raise
+        except Exception:
+            logger.exception("Failed to reserve quota for %s, feature %s", tenant_id, quota_type.value)
+            return unlimited()
+
+    @staticmethod
+    def check(quota_type: QuotaType, tenant_id: str, amount: int = 1) -> bool:
+        if not dify_config.BILLING_ENABLED:
+            return True
+
+        if amount <= 0:
+            raise ValueError("Amount to check must be greater than 0")
+
+        try:
+            remaining = QuotaService.get_remaining(quota_type, tenant_id)
+            return remaining >= amount if remaining != -1 else True
+        except Exception:
+            logger.exception("Failed to check quota for %s, feature %s", tenant_id, quota_type.value)
+            return True
+
+    @staticmethod
+    def release(quota_type: QuotaType, reservation_id: str, tenant_id: str, feature_key: str) -> None:
+        """Release a reservation. Guarantees no exceptions."""
+        try:
+            from services.billing_service import BillingService
+
+            if not dify_config.BILLING_ENABLED:
+                return
+
+            if not reservation_id:
+                return
+
+            logger.info("Releasing %s quota, reservation_id: %s", quota_type.value, reservation_id)
+            BillingService.quota_release(
+                tenant_id=tenant_id,
+                feature_key=feature_key,
+                reservation_id=reservation_id,
+            )
+        except Exception:
+            logger.exception("Failed to release quota, reservation_id: %s", reservation_id)
+
+    @staticmethod
+    def get_remaining(quota_type: QuotaType, tenant_id: str) -> int:
+        from services.billing_service import BillingService
+
+        try:
+            usage_info = BillingService.get_quota_info(tenant_id)
+            if isinstance(usage_info, dict):
+                feature_info = usage_info.get(quota_type.billing_key, {})
+                if isinstance(feature_info, dict):
+                    limit = feature_info.get("limit", 0)
+                    usage = feature_info.get("usage", 0)
+                    if limit == -1:
+                        return -1
+                    return max(0, limit - usage)
+            return 0
+        except Exception:
+            logger.exception("Failed to get remaining quota for %s, feature %s", tenant_id, quota_type.value)
+            return -1

api/services/trigger/webhook_service.py

@@ -38,6 +38,7 @@ from models.workflow import Workflow
 from services.async_workflow_service import AsyncWorkflowService
 from services.end_user_service import EndUserService
 from services.errors.app import QuotaExceededError
+from services.quota_service import QuotaService
 from services.trigger.app_trigger_service import AppTriggerService
 from services.workflow.entities import WebhookTriggerData
 
@@ -782,9 +783,9 @@ class WebhookService:
                     user_id=None,
                 )
 
-                # consume quota before triggering workflow execution
+                # reserve quota before triggering workflow execution
                 try:
-                    QuotaType.TRIGGER.consume(webhook_trigger.tenant_id)
+                    quota_charge = QuotaService.reserve(QuotaType.TRIGGER, webhook_trigger.tenant_id)
                 except QuotaExceededError:
                     AppTriggerService.mark_tenant_triggers_rate_limited(webhook_trigger.tenant_id)
                     logger.info(
@@ -795,11 +796,16 @@ class WebhookService:
                     raise
 
                 # Trigger workflow execution asynchronously
-                AsyncWorkflowService.trigger_workflow_async(
-                    session,
-                    end_user,
-                    trigger_data,
-                )
+                try:
+                    AsyncWorkflowService.trigger_workflow_async(
+                        session,
+                        end_user,
+                        trigger_data,
+                    )
+                    quota_charge.commit()
+                except Exception:
+                    quota_charge.refund()
+                    raise
 
         except Exception:
             logger.exception("Failed to trigger workflow for webhook %s", webhook_trigger.webhook_id)

api/tasks/trigger_processing_tasks.py

@@ -28,7 +28,7 @@ from core.trigger.entities.entities import TriggerProviderEntity
 from core.trigger.provider import PluginTriggerProviderController
 from core.trigger.trigger_manager import TriggerManager
 from core.workflow.nodes.trigger_plugin.entities import TriggerEventNodeData
-from enums.quota_type import QuotaType, unlimited
+from enums.quota_type import QuotaType
 from models.enums import (
     AppTriggerType,
     CreatorUserRole,
@@ -42,6 +42,7 @@ from models.workflow import Workflow, WorkflowAppLog, WorkflowAppLogCreatedFrom,
 from services.async_workflow_service import AsyncWorkflowService
 from services.end_user_service import EndUserService
 from services.errors.app import QuotaExceededError
+from services.quota_service import QuotaService, unlimited
 from services.trigger.app_trigger_service import AppTriggerService
 from services.trigger.trigger_provider_service import TriggerProviderService
 from services.trigger.trigger_request_service import TriggerHttpRequestCachingService
@@ -298,10 +299,10 @@ def dispatch_triggered_workflow(
                 icon_dark_filename=trigger_entity.identity.icon_dark or "",
             )
 
-            # consume quota before invoking trigger
+            # reserve quota before invoking trigger
             quota_charge = unlimited()
             try:
-                quota_charge = QuotaType.TRIGGER.consume(subscription.tenant_id)
+                quota_charge = QuotaService.reserve(QuotaType.TRIGGER, subscription.tenant_id)
             except QuotaExceededError:
                 AppTriggerService.mark_tenant_triggers_rate_limited(subscription.tenant_id)
                 logger.info(
@@ -387,6 +388,7 @@ def dispatch_triggered_workflow(
                     raise ValueError(f"End user not found for app {plugin_trigger.app_id}")
 
                 AsyncWorkflowService.trigger_workflow_async(session=session, user=end_user, trigger_data=trigger_data)
+                quota_charge.commit()
                 dispatched_count += 1
                 logger.info(
                     "Triggered workflow for app %s with trigger event %s",

api/tasks/workflow_schedule_tasks.py

@@ -8,10 +8,11 @@ from core.workflow.nodes.trigger_schedule.exc import (
     ScheduleNotFoundError,
     TenantOwnerNotFoundError,
 )
-from enums.quota_type import QuotaType, unlimited
+from enums.quota_type import QuotaType
 from models.trigger import WorkflowSchedulePlan
 from services.async_workflow_service import AsyncWorkflowService
 from services.errors.app import QuotaExceededError
+from services.quota_service import QuotaService, unlimited
 from services.trigger.app_trigger_service import AppTriggerService
 from services.trigger.schedule_service import ScheduleService
 from services.workflow.entities import ScheduleTriggerData
@@ -43,7 +44,7 @@ def run_schedule_trigger(schedule_id: str) -> None:
 
         quota_charge = unlimited()
         try:
-            quota_charge = QuotaType.TRIGGER.consume(schedule.tenant_id)
+            quota_charge = QuotaService.reserve(QuotaType.TRIGGER, schedule.tenant_id)
         except QuotaExceededError:
             AppTriggerService.mark_tenant_triggers_rate_limited(schedule.tenant_id)
             logger.info("Tenant %s rate limited, skipping schedule trigger %s", schedule.tenant_id, schedule_id)
@@ -61,6 +62,7 @@ def run_schedule_trigger(schedule_id: str) -> None:
                     tenant_id=schedule.tenant_id,
                 ),
             )
+            quota_charge.commit()
             logger.info("Schedule %s triggered workflow: %s", schedule_id, response.workflow_trigger_log_id)
         except Exception as e:
             quota_charge.refund()

api/tests/test_containers_integration_tests/services/test_app_generate_service.py

@@ -36,12 +36,19 @@ class TestAppGenerateService:
             ) as mock_message_based_generator,
             patch("services.account_service.FeatureService", autospec=True) as mock_account_feature_service,
             patch("services.app_generate_service.dify_config", autospec=True) as mock_dify_config,
+            patch("services.quota_service.dify_config", autospec=True) as mock_quota_dify_config,
             patch("configs.dify_config", autospec=True) as mock_global_dify_config,
         ):
             # Setup default mock returns for billing service
-            mock_billing_service.update_tenant_feature_plan_usage.return_value = {
-                "result": "success",
-                "history_id": "test_history_id",
+            mock_billing_service.quota_reserve.return_value = {
+                "reservation_id": "test-reservation-id",
+                "available": 100,
+                "reserved": 1,
+            }
+            mock_billing_service.quota_commit.return_value = {
+                "available": 99,
+                "reserved": 0,
+                "refunded": 0,
             }
 
             # Setup default mock returns for workflow service
@@ -101,6 +108,8 @@ class TestAppGenerateService:
             mock_dify_config.APP_DEFAULT_ACTIVE_REQUESTS = 100
             mock_dify_config.APP_DAILY_RATE_LIMIT = 1000
 
+            mock_quota_dify_config.BILLING_ENABLED = False
+
             mock_global_dify_config.BILLING_ENABLED = False
             mock_global_dify_config.APP_MAX_ACTIVE_REQUESTS = 100
             mock_global_dify_config.APP_DAILY_RATE_LIMIT = 1000
@@ -118,6 +127,7 @@ class TestAppGenerateService:
                 "message_based_generator": mock_message_based_generator,
                 "account_feature_service": mock_account_feature_service,
                 "dify_config": mock_dify_config,
+                "quota_dify_config": mock_quota_dify_config,
                 "global_dify_config": mock_global_dify_config,
             }
 
@@ -465,6 +475,7 @@ class TestAppGenerateService:
 
         # Set BILLING_ENABLED to True for this test
         mock_external_service_dependencies["dify_config"].BILLING_ENABLED = True
+        mock_external_service_dependencies["quota_dify_config"].BILLING_ENABLED = True
         mock_external_service_dependencies["global_dify_config"].BILLING_ENABLED = True
 
         # Setup test arguments
@@ -478,8 +489,10 @@ class TestAppGenerateService:
         # Verify the result
         assert result == ["test_response"]
 
-        # Verify billing service was called to consume quota
-        mock_external_service_dependencies["billing_service"].update_tenant_feature_plan_usage.assert_called_once()
+        # Verify billing two-phase quota (reserve + commit)
+        billing = mock_external_service_dependencies["billing_service"]
+        billing.quota_reserve.assert_called_once()
+        billing.quota_commit.assert_called_once()
 
     def test_generate_with_invalid_app_mode(
         self, db_session_with_containers: Session, mock_external_service_dependencies

api/tests/test_containers_integration_tests/trigger/test_trigger_e2e.py

@@ -602,9 +602,9 @@ def test_schedule_trigger_creates_trigger_log(
     )
 
     # Mock quota to avoid rate limiting
-    from enums import quota_type
+    from services import quota_service
 
-    monkeypatch.setattr(quota_type.QuotaType.TRIGGER, "consume", lambda _tenant_id: quota_type.unlimited())
+    monkeypatch.setattr(quota_service.QuotaService, "reserve", lambda *_args, **_kwargs: quota_service.unlimited())
 
     # Execute schedule trigger
     workflow_schedule_tasks.run_schedule_trigger(plan.id)

api/tests/unit_tests/enums/test_quota_type.py

@@ -0,0 +1,349 @@
+"""Unit tests for QuotaType, QuotaService, and QuotaCharge."""
+
+from unittest.mock import patch
+
+import pytest
+
+from enums.quota_type import QuotaType
+from services.quota_service import QuotaCharge, QuotaService, unlimited
+
+
+class TestQuotaType:
+    def test_billing_key_trigger(self):
+        assert QuotaType.TRIGGER.billing_key == "trigger_event"
+
+    def test_billing_key_workflow(self):
+        assert QuotaType.WORKFLOW.billing_key == "api_rate_limit"
+
+    def test_billing_key_unlimited_raises(self):
+        with pytest.raises(ValueError, match="Invalid quota type"):
+            _ = QuotaType.UNLIMITED.billing_key
+
+
+class TestQuotaService:
+    def test_reserve_billing_disabled(self):
+        with (
+            patch("services.quota_service.dify_config") as mock_cfg,
+            patch("services.billing_service.BillingService"),
+        ):
+            mock_cfg.BILLING_ENABLED = False
+            charge = QuotaService.reserve(QuotaType.TRIGGER, "t1")
+            assert charge.success is True
+            assert charge.charge_id is None
+
+    def test_reserve_zero_amount_raises(self):
+        with patch("services.quota_service.dify_config") as mock_cfg:
+            mock_cfg.BILLING_ENABLED = True
+            with pytest.raises(ValueError, match="greater than 0"):
+                QuotaService.reserve(QuotaType.TRIGGER, "t1", amount=0)
+
+    def test_reserve_success(self):
+        with (
+            patch("services.quota_service.dify_config") as mock_cfg,
+            patch("services.billing_service.BillingService") as mock_bs,
+        ):
+            mock_cfg.BILLING_ENABLED = True
+            mock_bs.quota_reserve.return_value = {"reservation_id": "rid-1", "available": 99}
+
+            charge = QuotaService.reserve(QuotaType.TRIGGER, "t1", amount=1)
+
+            assert charge.success is True
+            assert charge.charge_id == "rid-1"
+            assert charge._tenant_id == "t1"
+            assert charge._feature_key == "trigger_event"
+            assert charge._amount == 1
+            mock_bs.quota_reserve.assert_called_once()
+
+    def test_reserve_no_reservation_id_raises(self):
+        from services.errors.app import QuotaExceededError
+
+        with (
+            patch("services.quota_service.dify_config") as mock_cfg,
+            patch("services.billing_service.BillingService") as mock_bs,
+        ):
+            mock_cfg.BILLING_ENABLED = True
+            mock_bs.quota_reserve.return_value = {}
+
+            with pytest.raises(QuotaExceededError):
+                QuotaService.reserve(QuotaType.TRIGGER, "t1")
+
+    def test_reserve_quota_exceeded_propagates(self):
+        from services.errors.app import QuotaExceededError
+
+        with (
+            patch("services.quota_service.dify_config") as mock_cfg,
+            patch("services.billing_service.BillingService") as mock_bs,
+        ):
+            mock_cfg.BILLING_ENABLED = True
+            mock_bs.quota_reserve.side_effect = QuotaExceededError(feature="trigger", tenant_id="t1", required=1)
+
+            with pytest.raises(QuotaExceededError):
+                QuotaService.reserve(QuotaType.TRIGGER, "t1")
+
+    def test_reserve_api_exception_returns_unlimited(self):
+        with (
+            patch("services.quota_service.dify_config") as mock_cfg,
+            patch("services.billing_service.BillingService") as mock_bs,
+        ):
+            mock_cfg.BILLING_ENABLED = True
+            mock_bs.quota_reserve.side_effect = RuntimeError("network")
+
+            charge = QuotaService.reserve(QuotaType.TRIGGER, "t1")
+            assert charge.success is True
+            assert charge.charge_id is None
+
+    def test_consume_calls_reserve_and_commit(self):
+        with (
+            patch("services.quota_service.dify_config") as mock_cfg,
+            patch("services.billing_service.BillingService") as mock_bs,
+        ):
+            mock_cfg.BILLING_ENABLED = True
+            mock_bs.quota_reserve.return_value = {"reservation_id": "rid-c"}
+            mock_bs.quota_commit.return_value = {}
+
+            charge = QuotaService.consume(QuotaType.TRIGGER, "t1")
+            assert charge.success is True
+            mock_bs.quota_commit.assert_called_once()
+
+    def test_check_billing_disabled(self):
+        with patch("services.quota_service.dify_config") as mock_cfg:
+            mock_cfg.BILLING_ENABLED = False
+            assert QuotaService.check(QuotaType.TRIGGER, "t1") is True
+
+    def test_check_zero_amount_raises(self):
+        with patch("services.quota_service.dify_config") as mock_cfg:
+            mock_cfg.BILLING_ENABLED = True
+            with pytest.raises(ValueError, match="greater than 0"):
+                QuotaService.check(QuotaType.TRIGGER, "t1", amount=0)
+
+    def test_check_sufficient_quota(self):
+        with (
+            patch("services.quota_service.dify_config") as mock_cfg,
+            patch.object(QuotaService, "get_remaining", return_value=100),
+        ):
+            mock_cfg.BILLING_ENABLED = True
+            assert QuotaService.check(QuotaType.TRIGGER, "t1", amount=50) is True
+
+    def test_check_insufficient_quota(self):
+        with (
+            patch("services.quota_service.dify_config") as mock_cfg,
+            patch.object(QuotaService, "get_remaining", return_value=5),
+        ):
+            mock_cfg.BILLING_ENABLED = True
+            assert QuotaService.check(QuotaType.TRIGGER, "t1", amount=10) is False
+
+    def test_check_unlimited_quota(self):
+        with (
+            patch("services.quota_service.dify_config") as mock_cfg,
+            patch.object(QuotaService, "get_remaining", return_value=-1),
+        ):
+            mock_cfg.BILLING_ENABLED = True
+            assert QuotaService.check(QuotaType.TRIGGER, "t1", amount=999) is True
+
+    def test_check_exception_returns_true(self):
+        with (
+            patch("services.quota_service.dify_config") as mock_cfg,
+            patch.object(QuotaService, "get_remaining", side_effect=RuntimeError),
+        ):
+            mock_cfg.BILLING_ENABLED = True
+            assert QuotaService.check(QuotaType.TRIGGER, "t1") is True
+
+    def test_release_billing_disabled(self):
+        with (
+            patch("services.quota_service.dify_config") as mock_cfg,
+            patch("services.billing_service.BillingService") as mock_bs,
+        ):
+            mock_cfg.BILLING_ENABLED = False
+            QuotaService.release(QuotaType.TRIGGER, "rid-1", "t1", "trigger_event")
+            mock_bs.quota_release.assert_not_called()
+
+    def test_release_empty_reservation(self):
+        with (
+            patch("services.quota_service.dify_config") as mock_cfg,
+            patch("services.billing_service.BillingService") as mock_bs,
+        ):
+            mock_cfg.BILLING_ENABLED = True
+            QuotaService.release(QuotaType.TRIGGER, "", "t1", "trigger_event")
+            mock_bs.quota_release.assert_not_called()
+
+    def test_release_success(self):
+        with (
+            patch("services.quota_service.dify_config") as mock_cfg,
+            patch("services.billing_service.BillingService") as mock_bs,
+        ):
+            mock_cfg.BILLING_ENABLED = True
+            mock_bs.quota_release.return_value = {}
+            QuotaService.release(QuotaType.TRIGGER, "rid-1", "t1", "trigger_event")
+            mock_bs.quota_release.assert_called_once_with(
+                tenant_id="t1", feature_key="trigger_event", reservation_id="rid-1"
+            )
+
+    def test_release_exception_swallowed(self):
+        with (
+            patch("services.quota_service.dify_config") as mock_cfg,
+            patch("services.billing_service.BillingService") as mock_bs,
+        ):
+            mock_cfg.BILLING_ENABLED = True
+            mock_bs.quota_release.side_effect = RuntimeError("fail")
+            QuotaService.release(QuotaType.TRIGGER, "rid-1", "t1", "trigger_event")
+
+    def test_get_remaining_normal(self):
+        with patch("services.billing_service.BillingService") as mock_bs:
+            mock_bs.get_quota_info.return_value = {"trigger_event": {"limit": 100, "usage": 30}}
+            assert QuotaService.get_remaining(QuotaType.TRIGGER, "t1") == 70
+
+    def test_get_remaining_unlimited(self):
+        with patch("services.billing_service.BillingService") as mock_bs:
+            mock_bs.get_quota_info.return_value = {"trigger_event": {"limit": -1, "usage": 0}}
+            assert QuotaService.get_remaining(QuotaType.TRIGGER, "t1") == -1
+
+    def test_get_remaining_over_limit_returns_zero(self):
+        with patch("services.billing_service.BillingService") as mock_bs:
+            mock_bs.get_quota_info.return_value = {"trigger_event": {"limit": 10, "usage": 15}}
+            assert QuotaService.get_remaining(QuotaType.TRIGGER, "t1") == 0
+
+    def test_get_remaining_exception_returns_neg1(self):
+        with patch("services.billing_service.BillingService") as mock_bs:
+            mock_bs.get_quota_info.side_effect = RuntimeError
+            assert QuotaService.get_remaining(QuotaType.TRIGGER, "t1") == -1
+
+    def test_get_remaining_empty_response(self):
+        with patch("services.billing_service.BillingService") as mock_bs:
+            mock_bs.get_quota_info.return_value = {}
+            assert QuotaService.get_remaining(QuotaType.TRIGGER, "t1") == 0
+
+    def test_get_remaining_non_dict_response(self):
+        with patch("services.billing_service.BillingService") as mock_bs:
+            mock_bs.get_quota_info.return_value = "invalid"
+            assert QuotaService.get_remaining(QuotaType.TRIGGER, "t1") == 0
+
+    def test_get_remaining_feature_not_in_response(self):
+        with patch("services.billing_service.BillingService") as mock_bs:
+            mock_bs.get_quota_info.return_value = {"other_feature": {"limit": 100, "usage": 0}}
+            remaining = QuotaService.get_remaining(QuotaType.TRIGGER, "t1")
+            assert remaining == 0
+
+    def test_get_remaining_non_dict_feature_info(self):
+        with patch("services.billing_service.BillingService") as mock_bs:
+            mock_bs.get_quota_info.return_value = {"trigger_event": "not_a_dict"}
+            assert QuotaService.get_remaining(QuotaType.TRIGGER, "t1") == 0
+
+
+class TestQuotaCharge:
+    def test_commit_success(self):
+        with patch("services.billing_service.BillingService") as mock_bs:
+            mock_bs.quota_commit.return_value = {}
+            charge = QuotaCharge(
+                success=True,
+                charge_id="rid-1",
+                _quota_type=QuotaType.TRIGGER,
+                _tenant_id="t1",
+                _feature_key="trigger_event",
+                _amount=1,
+            )
+            charge.commit()
+            mock_bs.quota_commit.assert_called_once_with(
+                tenant_id="t1",
+                feature_key="trigger_event",
+                reservation_id="rid-1",
+                actual_amount=1,
+            )
+            assert charge._committed is True
+
+    def test_commit_with_actual_amount(self):
+        with patch("services.billing_service.BillingService") as mock_bs:
+            mock_bs.quota_commit.return_value = {}
+            charge = QuotaCharge(
+                success=True,
+                charge_id="rid-1",
+                _quota_type=QuotaType.TRIGGER,
+                _tenant_id="t1",
+                _feature_key="trigger_event",
+                _amount=10,
+            )
+            charge.commit(actual_amount=5)
+            call_kwargs = mock_bs.quota_commit.call_args[1]
+            assert call_kwargs["actual_amount"] == 5
+
+    def test_commit_idempotent(self):
+        with patch("services.billing_service.BillingService") as mock_bs:
+            mock_bs.quota_commit.return_value = {}
+            charge = QuotaCharge(
+                success=True,
+                charge_id="rid-1",
+                _quota_type=QuotaType.TRIGGER,
+                _tenant_id="t1",
+                _feature_key="trigger_event",
+                _amount=1,
+            )
+            charge.commit()
+            charge.commit()
+            assert mock_bs.quota_commit.call_count == 1
+
+    def test_commit_no_charge_id_noop(self):
+        with patch("services.billing_service.BillingService") as mock_bs:
+            charge = QuotaCharge(success=True, charge_id=None, _quota_type=QuotaType.TRIGGER)
+            charge.commit()
+            mock_bs.quota_commit.assert_not_called()
+
+    def test_commit_no_tenant_id_noop(self):
+        with patch("services.billing_service.BillingService") as mock_bs:
+            charge = QuotaCharge(
+                success=True,
+                charge_id="rid-1",
+                _quota_type=QuotaType.TRIGGER,
+                _tenant_id=None,
+                _feature_key="trigger_event",
+            )
+            charge.commit()
+            mock_bs.quota_commit.assert_not_called()
+
+    def test_commit_exception_swallowed(self):
+        with patch("services.billing_service.BillingService") as mock_bs:
+            mock_bs.quota_commit.side_effect = RuntimeError("fail")
+            charge = QuotaCharge(
+                success=True,
+                charge_id="rid-1",
+                _quota_type=QuotaType.TRIGGER,
+                _tenant_id="t1",
+                _feature_key="trigger_event",
+                _amount=1,
+            )
+            charge.commit()
+
+    def test_refund_success(self):
+        with patch.object(QuotaService, "release") as mock_rel:
+            charge = QuotaCharge(
+                success=True,
+                charge_id="rid-1",
+                _quota_type=QuotaType.TRIGGER,
+                _tenant_id="t1",
+                _feature_key="trigger_event",
+            )
+            charge.refund()
+            mock_rel.assert_called_once_with(QuotaType.TRIGGER, "rid-1", "t1", "trigger_event")
+
+    def test_refund_no_charge_id_noop(self):
+        with patch.object(QuotaService, "release") as mock_rel:
+            charge = QuotaCharge(success=True, charge_id=None, _quota_type=QuotaType.TRIGGER)
+            charge.refund()
+            mock_rel.assert_not_called()
+
+    def test_refund_no_tenant_id_noop(self):
+        with patch.object(QuotaService, "release") as mock_rel:
+            charge = QuotaCharge(
+                success=True,
+                charge_id="rid-1",
+                _quota_type=QuotaType.TRIGGER,
+                _tenant_id=None,
+            )
+            charge.refund()
+            mock_rel.assert_not_called()
+
+
+class TestUnlimited:
+    def test_unlimited_returns_success_with_no_charge_id(self):
+        charge = unlimited()
+        assert charge.success is True
+        assert charge.charge_id is None
+        assert charge._quota_type == QuotaType.UNLIMITED

api/tests/unit_tests/services/test_app_generate_service.py

@@ -23,6 +23,7 @@ import pytest
 
 import services.app_generate_service as ags_module
 from core.app.entities.app_invoke_entities import InvokeFrom
+from enums.quota_type import QuotaType
 from models.model import AppMode
 from services.app_generate_service import AppGenerateService
 from services.errors.app import WorkflowIdFormatError, WorkflowNotFoundError
@@ -447,8 +448,8 @@ class TestGenerateBilling:
     def test_billing_enabled_consumes_quota(self, mocker, monkeypatch):
         monkeypatch.setattr(ags_module.dify_config, "BILLING_ENABLED", True)
         quota_charge = MagicMock()
-        consume_mock = mocker.patch(
-            "services.app_generate_service.QuotaType.WORKFLOW.consume",
+        reserve_mock = mocker.patch(
+            "services.app_generate_service.QuotaService.reserve",
             return_value=quota_charge,
         )
         mocker.patch(
@@ -467,7 +468,8 @@ class TestGenerateBilling:
             invoke_from=InvokeFrom.SERVICE_API,
             streaming=False,
         )
-        consume_mock.assert_called_once_with("tenant-id")
+        reserve_mock.assert_called_once_with(QuotaType.WORKFLOW, "tenant-id")
+        quota_charge.commit.assert_called_once()
 
     def test_billing_quota_exceeded_raises_rate_limit_error(self, mocker, monkeypatch):
         from services.errors.app import QuotaExceededError
@@ -475,7 +477,7 @@ class TestGenerateBilling:
 
         monkeypatch.setattr(ags_module.dify_config, "BILLING_ENABLED", True)
         mocker.patch(
-            "services.app_generate_service.QuotaType.WORKFLOW.consume",
+            "services.app_generate_service.QuotaService.reserve",
             side_effect=QuotaExceededError(feature="workflow", tenant_id="t", required=1),
         )
 
@@ -492,7 +494,7 @@ class TestGenerateBilling:
         monkeypatch.setattr(ags_module.dify_config, "BILLING_ENABLED", True)
         quota_charge = MagicMock()
         mocker.patch(
-            "services.app_generate_service.QuotaType.WORKFLOW.consume",
+            "services.app_generate_service.QuotaService.reserve",
             return_value=quota_charge,
         )
         mocker.patch(

api/tests/unit_tests/services/test_async_workflow_service.py

@@ -57,7 +57,7 @@ class TestAsyncWorkflowService:
             - repo: SQLAlchemyWorkflowTriggerLogRepository
             - dispatcher_manager_class: QueueDispatcherManager class
             - dispatcher: dispatcher instance
-            - quota_workflow: QuotaType.WORKFLOW
+            - quota_service: QuotaService mock
             - get_workflow: AsyncWorkflowService._get_workflow method
             - professional_task: execute_workflow_professional
             - team_task: execute_workflow_team
@@ -72,7 +72,7 @@ class TestAsyncWorkflowService:
         mock_repo.create.side_effect = _create_side_effect
 
         mock_dispatcher = MagicMock()
-        quota_workflow = MagicMock()
+        mock_quota_service = MagicMock()
         mock_get_workflow = MagicMock()
 
         mock_professional_task = MagicMock()
@@ -93,8 +93,8 @@ class TestAsyncWorkflowService:
             ) as mock_get_workflow,
             patch.object(
                 async_workflow_service_module,
-                "QuotaType",
-                new=SimpleNamespace(WORKFLOW=quota_workflow),
+                "QuotaService",
+                new=mock_quota_service,
             ),
             patch.object(async_workflow_service_module, "execute_workflow_professional") as mock_professional_task,
             patch.object(async_workflow_service_module, "execute_workflow_team") as mock_team_task,
@@ -107,7 +107,7 @@ class TestAsyncWorkflowService:
                 "repo": mock_repo,
                 "dispatcher_manager_class": mock_dispatcher_manager_class,
                 "dispatcher": mock_dispatcher,
-                "quota_workflow": quota_workflow,
+                "quota_service": mock_quota_service,
                 "get_workflow": mock_get_workflow,
                 "professional_task": mock_professional_task,
                 "team_task": mock_team_task,
@@ -146,6 +146,9 @@ class TestAsyncWorkflowService:
         mocks["team_task"].delay.return_value = task_result
         mocks["sandbox_task"].delay.return_value = task_result
 
+        quota_charge_mock = MagicMock()
+        mocks["quota_service"].reserve.return_value = quota_charge_mock
+
         class DummyAccount:
             def __init__(self, user_id: str):
                 self.id = user_id
@@ -163,7 +166,8 @@ class TestAsyncWorkflowService:
         assert result.status == "queued"
         assert result.queue == queue_name
 
-        mocks["quota_workflow"].consume.assert_called_once_with("tenant-123")
+        mocks["quota_service"].reserve.assert_called_once()
+        quota_charge_mock.commit.assert_called_once()
         assert session.commit.call_count == 2
 
         created_log = mocks["repo"].create.call_args[0][0]
@@ -250,7 +254,7 @@ class TestAsyncWorkflowService:
         mocks = async_workflow_trigger_mocks
         mocks["dispatcher"].get_queue_name.return_value = QueuePriority.TEAM
         mocks["get_workflow"].return_value = workflow
-        mocks["quota_workflow"].consume.side_effect = QuotaExceededError(
+        mocks["quota_service"].reserve.side_effect = QuotaExceededError(
             feature="workflow",
             tenant_id="tenant-123",
             required=1,

api/tests/unit_tests/services/test_billing_service.py

@@ -415,7 +415,7 @@ class TestBillingServiceUsageCalculation:
             yield mock
 
     def test_get_tenant_feature_plan_usage_info(self, mock_send_request):
-        """Test retrieval of tenant feature plan usage information."""
+        """Test retrieval of tenant feature plan usage information (legacy endpoint)."""
         # Arrange
         tenant_id = "tenant-123"
         expected_response = {"features": {"trigger": {"used": 50, "limit": 100}, "workflow": {"used": 20, "limit": 50}}}
@@ -428,6 +428,20 @@ class TestBillingServiceUsageCalculation:
         assert result == expected_response
         mock_send_request.assert_called_once_with("GET", "/tenant-feature-usage/info", params={"tenant_id": tenant_id})
 
+    def test_get_quota_info(self, mock_send_request):
+        """Test retrieval of quota info from new endpoint."""
+        # Arrange
+        tenant_id = "tenant-123"
+        expected_response = {"trigger_event": {"limit": 100, "usage": 30}, "api_rate_limit": {"limit": -1, "usage": 0}}
+        mock_send_request.return_value = expected_response
+
+        # Act
+        result = BillingService.get_quota_info(tenant_id)
+
+        # Assert
+        assert result == expected_response
+        mock_send_request.assert_called_once_with("GET", "/quota/info", params={"tenant_id": tenant_id})
+
     def test_update_tenant_feature_plan_usage_positive_delta(self, mock_send_request):
         """Test updating tenant feature usage with positive delta (adding credits)."""
         # Arrange
@@ -505,6 +519,118 @@ class TestBillingServiceUsageCalculation:
         )
 
 
+class TestBillingServiceQuotaOperations:
+    """Unit tests for quota reserve/commit/release operations."""
+
+    @pytest.fixture
+    def mock_send_request(self):
+        with patch.object(BillingService, "_send_request") as mock:
+            yield mock
+
+    def test_quota_reserve_success(self, mock_send_request):
+        expected = {"reservation_id": "rid-1", "available": 99, "reserved": 1}
+        mock_send_request.return_value = expected
+
+        result = BillingService.quota_reserve(tenant_id="t1", feature_key="trigger_event", request_id="req-1", amount=1)
+
+        assert result == expected
+        mock_send_request.assert_called_once_with(
+            "POST",
+            "/quota/reserve",
+            json={"tenant_id": "t1", "feature_key": "trigger_event", "request_id": "req-1", "amount": 1},
+        )
+
+    def test_quota_reserve_coerces_string_to_int(self, mock_send_request):
+        """Test that TypeAdapter coerces string values to int."""
+        mock_send_request.return_value = {"reservation_id": "rid-str", "available": "99", "reserved": "1"}
+
+        result = BillingService.quota_reserve(tenant_id="t1", feature_key="trigger_event", request_id="req-s", amount=1)
+
+        assert result["available"] == 99
+        assert isinstance(result["available"], int)
+        assert result["reserved"] == 1
+        assert isinstance(result["reserved"], int)
+
+    def test_quota_reserve_with_meta(self, mock_send_request):
+        mock_send_request.return_value = {"reservation_id": "rid-2", "available": 98, "reserved": 1}
+        meta = {"source": "webhook"}
+
+        BillingService.quota_reserve(
+            tenant_id="t1", feature_key="trigger_event", request_id="req-2", amount=1, meta=meta
+        )
+
+        call_json = mock_send_request.call_args[1]["json"]
+        assert call_json["meta"] == {"source": "webhook"}
+
+    def test_quota_commit_success(self, mock_send_request):
+        expected = {"available": 98, "reserved": 0, "refunded": 0}
+        mock_send_request.return_value = expected
+
+        result = BillingService.quota_commit(
+            tenant_id="t1", feature_key="trigger_event", reservation_id="rid-1", actual_amount=1
+        )
+
+        assert result == expected
+        mock_send_request.assert_called_once_with(
+            "POST",
+            "/quota/commit",
+            json={
+                "tenant_id": "t1",
+                "feature_key": "trigger_event",
+                "reservation_id": "rid-1",
+                "actual_amount": 1,
+            },
+        )
+
+    def test_quota_commit_coerces_string_to_int(self, mock_send_request):
+        """Test that TypeAdapter coerces string values to int."""
+        mock_send_request.return_value = {"available": "97", "reserved": "0", "refunded": "1"}
+
+        result = BillingService.quota_commit(
+            tenant_id="t1", feature_key="trigger_event", reservation_id="rid-s", actual_amount=1
+        )
+
+        assert result["available"] == 97
+        assert isinstance(result["available"], int)
+        assert result["refunded"] == 1
+        assert isinstance(result["refunded"], int)
+
+    def test_quota_commit_with_meta(self, mock_send_request):
+        mock_send_request.return_value = {"available": 97, "reserved": 0, "refunded": 0}
+        meta = {"reason": "partial"}
+
+        BillingService.quota_commit(
+            tenant_id="t1", feature_key="trigger_event", reservation_id="rid-1", actual_amount=1, meta=meta
+        )
+
+        call_json = mock_send_request.call_args[1]["json"]
+        assert call_json["meta"] == {"reason": "partial"}
+
+    def test_quota_release_success(self, mock_send_request):
+        expected = {"available": 100, "reserved": 0, "released": 1}
+        mock_send_request.return_value = expected
+
+        result = BillingService.quota_release(tenant_id="t1", feature_key="trigger_event", reservation_id="rid-1")
+
+        assert result == expected
+        mock_send_request.assert_called_once_with(
+            "POST",
+            "/quota/release",
+            json={"tenant_id": "t1", "feature_key": "trigger_event", "reservation_id": "rid-1"},
+        )
+
+    def test_quota_release_coerces_string_to_int(self, mock_send_request):
+        """Test that TypeAdapter coerces string values to int."""
+        mock_send_request.return_value = {"available": "100", "reserved": "0", "released": "1"}
+
+        result = BillingService.quota_release(tenant_id="t1", feature_key="trigger_event", reservation_id="rid-s")
+
+        assert result["available"] == 100
+        assert isinstance(result["available"], int)
+        assert result["released"] == 1
+        assert isinstance(result["released"], int)
+
+
 class TestBillingServiceRateLimitEnforcement:
     """Unit tests for rate limit enforcement mechanisms.