feat(cloud): enforce workspace resource quotas

This commit is contained in:
dadachann
2026-07-30 18:48:42 +00:00
parent 88f328066b
commit 59db012594
10 changed files with 722 additions and 12 deletions
@@ -14,6 +14,7 @@ from ....utils import bounded_executor
from ....workspace.collaboration import MembershipPermissionError, WorkspaceCollaborationError
from ....workspace.errors import WorkspaceNotFoundError
from ....cloud.entitlements import EntitlementUnavailableError
from ....cloud.quotas import WorkspaceQuotaExceededError
from ....core.errors import TaskCapacityError
from ..authz import AuthorizationError, Permission, permissions_for_role, require_permission
from ..context import PrincipalContext, PrincipalType, RequestContext, WorkspaceContext
@@ -219,6 +220,8 @@ class RouterGroup(abc.ABC):
return self.http_status(403, e.code, str(e))
if isinstance(e, WorkspaceCollaborationError):
return self.http_status(400, e.code, str(e))
if isinstance(e, WorkspaceQuotaExceededError):
return self.http_status(409, e.error_code, str(e))
if isinstance(e, TaskCapacityError):
return self.http_status(429, 'task_capacity_exceeded', str(e))
if isinstance(
+28 -8
View File
@@ -4,6 +4,7 @@ import uuid
import sqlalchemy
from ....core import app
from ....cloud.quotas import require_resource_capacity, resolve_workspace_quota
from ....entity.persistence import bot as persistence_bot
from ....entity.persistence import pipeline as persistence_pipeline
from ....workspace.errors import WorkspaceNotFoundError
@@ -101,20 +102,21 @@ class BotService:
async def create_bot(self, context: TenantContext, bot_data: dict) -> str:
"""Create bot"""
workspace_uuid = require_workspace_uuid(context)
# Check limitation
limitation = self.ap.instance_config.data.get('system', {}).get('limitation', {})
max_bots = limitation.get('max_bots', -1)
if max_bots >= 0:
existing_bots = await self.get_bots(context)
if len(existing_bots) >= max_bots:
raise ValueError(f'Maximum number of bots ({max_bots}) reached')
quota = await resolve_workspace_quota(
self.ap,
workspace_uuid,
'bots.max',
fallback=limitation.get('max_bots', -1),
)
# TODO: 检查配置信息格式
bot_data = bot_data.copy()
bot_data['uuid'] = str(uuid.uuid4())
bot_data['workspace_uuid'] = workspace_uuid
# bind the most recently updated pipeline if any exist
# Preserve the legacy flat-row result shape for this optional lookup;
# quota admission and insertion below still share one transaction.
result = await self.ap.persistence_mgr.execute_async(
scope_statement(
sqlalchemy.select(persistence_pipeline.LegacyPipeline),
@@ -129,7 +131,25 @@ class BotService:
bot_data['use_pipeline_uuid'] = pipeline.uuid
bot_data['use_pipeline_name'] = pipeline.name
await self.ap.persistence_mgr.execute_async(sqlalchemy.insert(persistence_bot.Bot).values(bot_data))
async def persist(execute) -> None:
await require_resource_capacity(
execute,
workspace_uuid=workspace_uuid,
model=persistence_bot.Bot,
quota=quota,
resource_name='bots',
)
await execute(sqlalchemy.insert(persistence_bot.Bot).values(bot_data))
tenant_uow = getattr(self.ap.persistence_mgr, 'tenant_uow', None)
if quota.requires_transaction_lock:
if not callable(tenant_uow):
raise RuntimeError('Cloud bot quota enforcement requires transactional persistence')
async with tenant_uow(workspace_uuid) as uow:
await persist(uow.execute)
else:
await persist(self.ap.persistence_mgr.execute_async)
bot = await self.get_bot(context, bot_data['uuid'], include_secret=True)
+82
View File
@@ -0,0 +1,82 @@
from __future__ import annotations
from dataclasses import dataclass
from typing import Any, Awaitable, Callable
import sqlalchemy
from ..entity.persistence import workspace as persistence_workspace
from .entitlements import EntitlementResolver
Execute = Callable[[Any], Awaitable[Any]]
class WorkspaceQuotaExceededError(ValueError):
"""A stable business error raised when a workspace has no free slots."""
error_code = 'workspace_quota_exceeded'
def __init__(self, resource_name: str, limit: int) -> None:
self.resource_name = resource_name
self.limit = limit
super().__init__(f'Maximum number of {resource_name} ({limit}) reached')
@dataclass(frozen=True, slots=True)
class WorkspaceQuota:
limit: int
requires_transaction_lock: bool
async def resolve_workspace_quota(
ap: Any,
workspace_uuid: str,
limit_name: str,
*,
fallback: int = -1,
) -> WorkspaceQuota:
"""Resolve a plan-agnostic Cloud limit while preserving OSS configuration."""
resolver = getattr(ap, 'entitlement_resolver', None)
if isinstance(resolver, EntitlementResolver):
snapshot = await resolver.resolve(workspace_uuid)
return WorkspaceQuota(
limit=snapshot.limit(limit_name),
requires_transaction_lock=True,
)
return WorkspaceQuota(limit=fallback, requires_transaction_lock=False)
async def lock_workspace_for_quota(execute: Execute, workspace_uuid: str) -> None:
"""Serialize quota checks on the durable Workspace row within one transaction."""
result = await execute(
sqlalchemy.select(persistence_workspace.Workspace.uuid)
.where(persistence_workspace.Workspace.uuid == workspace_uuid)
.with_for_update()
)
if result.first() is None:
raise ValueError('Workspace does not exist')
async def require_resource_capacity(
execute: Execute,
*,
workspace_uuid: str,
model: type,
quota: WorkspaceQuota,
resource_name: str,
workspace_locked: bool = False,
) -> None:
if quota.limit < 0:
return
if quota.requires_transaction_lock and not workspace_locked:
await lock_workspace_for_quota(execute, workspace_uuid)
result = await execute(
sqlalchemy.select(sqlalchemy.func.count())
.select_from(model)
.where(model.workspace_uuid == workspace_uuid)
)
if int(result.scalar_one()) >= quota.limit:
raise WorkspaceQuotaExceededError(resource_name, quota.limit)
+22
View File
@@ -19,6 +19,11 @@ from urllib.parse import urljoin, urlparse
from langbot_plugin.api.entities.builtin.pipeline.query import provider_session
from ..core import app
from ..cloud.quotas import (
lock_workspace_for_quota,
require_resource_capacity,
resolve_workspace_quota,
)
from . import handler
from .archive import inspect_plugin_archive_metadata
from .github import (
@@ -1295,6 +1300,11 @@ class PluginRuntimeConnector(ManagedRuntimeConnector):
install_info: dict[str, Any],
artifact_digest: str,
) -> tuple[InstallationBinding, str | None, bool]:
quota = await resolve_workspace_quota(
self.ap,
execution_context.workspace_uuid,
'plugins.max',
)
safe_install_info = {
key: value
for key, value in install_info.items()
@@ -1316,9 +1326,19 @@ class PluginRuntimeConnector(ManagedRuntimeConnector):
)
async def persist(execute):
if quota.requires_transaction_lock:
await lock_workspace_for_quota(execute, execution_context.workspace_uuid)
result = await execute(statement)
setting = result.first()
if setting is None:
await require_resource_capacity(
execute,
workspace_uuid=execution_context.workspace_uuid,
model=persistence_plugin.PluginSetting,
quota=quota,
resource_name='plugins',
workspace_locked=quota.requires_transaction_lock,
)
installation_uuid = str(uuid.uuid4())
runtime_revision = 1
previous_digest = None
@@ -1373,6 +1393,8 @@ class PluginRuntimeConnector(ManagedRuntimeConnector):
)
tenant_uow = getattr(self.ap.persistence_mgr, 'tenant_uow', None)
if quota.requires_transaction_lock and not callable(tenant_uow):
raise RuntimeError('Cloud plugin quota enforcement requires transactional persistence')
if callable(tenant_uow):
async with tenant_uow(execution_context.workspace_uuid) as uow:
return await persist(uow.execute)