aury-boot 0.0.29__tar.gz → 0.0.31__tar.gz

{aury_boot-0.0.29 → aury_boot-0.0.31}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: aury-boot
-Version: 0.0.29
+Version: 0.0.31
 Summary: Aury Boot - 基于 FastAPI 生态的企业级 API 开发框架
 Requires-Python: >=3.13
 Requires-Dist: alembic>=1.17.2
@@ -42,6 +42,14 @@ Requires-Dist: mkdocs-material>=9.7.0; extra == 'docs'
 Requires-Dist: mkdocs>=1.6.0; extra == 'docs'
 Provides-Extra: mysql
 Requires-Dist: aiomysql>=0.3.2; extra == 'mysql'
+Provides-Extra: otel
+Requires-Dist: opentelemetry-api>=1.25.0; extra == 'otel'
+Requires-Dist: opentelemetry-instrumentation-fastapi>=0.46b0; extra == 'otel'
+Requires-Dist: opentelemetry-instrumentation-httpx>=0.46b0; extra == 'otel'
+Requires-Dist: opentelemetry-instrumentation-sqlalchemy>=0.46b0; extra == 'otel'
+Requires-Dist: opentelemetry-sdk>=1.25.0; extra == 'otel'
+Provides-Extra: otel-exporter
+Requires-Dist: opentelemetry-exporter-otlp>=1.25.0; extra == 'otel-exporter'
 Provides-Extra: postgres
 Requires-Dist: asyncpg>=0.31.0; extra == 'postgres'
 Provides-Extra: rabbitmq
@@ -52,6 +60,11 @@ Requires-Dist: apscheduler>=3.11.1; extra == 'recommended'
 Requires-Dist: asyncpg>=0.31.0; extra == 'recommended'
 Requires-Dist: aury-sdk-storage[aws]>=0.0.1; extra == 'recommended'
 Requires-Dist: dramatiq>=1.18.0; extra == 'recommended'
+Requires-Dist: opentelemetry-api>=1.25.0; extra == 'recommended'
+Requires-Dist: opentelemetry-instrumentation-fastapi>=0.46b0; extra == 'recommended'
+Requires-Dist: opentelemetry-instrumentation-httpx>=0.46b0; extra == 'recommended'
+Requires-Dist: opentelemetry-instrumentation-sqlalchemy>=0.46b0; extra == 'recommended'
+Requires-Dist: opentelemetry-sdk>=1.25.0; extra == 'recommended'
 Requires-Dist: redis>=7.1.0; extra == 'recommended'
 Provides-Extra: redis
 Requires-Dist: redis>=7.1.0; extra == 'redis'

{aury_boot-0.0.29 → aury_boot-0.0.31}/aury/boot/_version.py

@@ -28,7 +28,7 @@ version_tuple: VERSION_TUPLE
 commit_id: COMMIT_ID
 __commit_id__: COMMIT_ID
 
-__version__ = version = '0.0.29'
-__version_tuple__ = version_tuple = (0, 0, 29)
+__version__ = version = '0.0.31'
+__version_tuple__ = version_tuple = (0, 0, 31)
 
 __commit_id__ = commit_id = None

{aury_boot-0.0.29 → aury_boot-0.0.31}/aury/boot/application/__init__.py

@@ -24,6 +24,7 @@ from aury.boot.domain.transaction import (
 # 依赖注入容器（从 infrastructure 导入）
 from aury.boot.infrastructure.di import Container, Lifetime, Scope, ServiceDescriptor
 from aury.boot.infrastructure.events import (
+    BroadcasterEventBus,
     Event,
     EventBackend,
     EventBusManager,
@@ -31,9 +32,7 @@ from aury.boot.infrastructure.events import (
     EventType,
     IEventBus,
     # 后端实现
-    MemoryEventBus,
     RabbitMQEventBus,
-    RedisEventBus,
 )
 
 from . import interfaces, rpc
@@ -108,14 +107,13 @@ __all__ = [
     "IEventBus",
     "Lifetime",
     "LogSettings",
-    "MemoryEventBus",
+    "BroadcasterEventBus",
     "Middleware",
     "MiddlewareName",
     "MigrationComponent",
     # 迁移
     "MigrationManager",
     "RabbitMQEventBus",
-    "RedisEventBus",
     "RequestLoggingMiddleware",
     "SchedulerComponent",
     "SchedulerMode",

{aury_boot-0.0.29 → aury_boot-0.0.31}/aury/boot/application/app/base.py

@@ -83,6 +83,78 @@ class Middleware(ABC):
         pass
 
 
+# TODO: 未来考虑合并 Middleware/Plugin/Component 为统一的 Extension 概念：
+# class Extension:
+#     def build_middleware(self, config) -> StarletteMiddleware | None  # 可选：app 创建前
+#     def install(self, app, config) -> None                           # 可选：app 创建后
+#     async def setup(self, app, config) -> None                       # 可选：lifespan 启动
+#     async def teardown(self, app) -> None                            # 可选：lifespan 关闭
+# 这样一个扩展可以同时实现多个 hook，适用于 OTEL、认证等复杂场景。
+
+
+class Plugin(ABC):
+    """应用插件基类。
+
+    用于需要 app 实例的扩展，如 OpenTelemetry instrument、Admin Console 挂载等。
+    插件在 app 创建后、lifespan 之前同步执行。
+
+    生命周期：
+    1. ``install()`` - 在 app 创建后同步调用，可以访问 app 实例。
+
+    与 Middleware 的区别：
+    - Middleware: 在 app 创建前构建，返回 Starlette Middleware
+    - Plugin: 在 app 创建后执行，可以操作 app 实例
+
+    与 Component 的区别：
+    - Component: 在 lifespan 里异步执行，用于基础设施初始化
+    - Plugin: 在 lifespan 前同步执行，用于 app 扩展
+
+    使用示例:
+        class TelemetryPlugin(Plugin):
+            name = "telemetry"
+            enabled = True
+
+            def install(self, app: FoundationApp, config: BaseConfig) -> None:
+                FastAPIInstrumentor.instrument_app(app)
+    """
+
+    name: str = "plugin"
+    enabled: bool = True
+
+    def can_enable(self, config: BaseConfig) -> bool:
+        """是否可以启用此插件。
+
+        子类可以重写此方法以实现条件化启用。
+
+        Args:
+            config: 应用配置
+
+        Returns:
+            是否启用
+        """
+        return self.enabled
+
+    @abstractmethod
+    def install(self, app: FoundationApp, config: BaseConfig) -> None:
+        """安装插件（同步，在 app 创建后、lifespan 之前调用）。
+
+        Args:
+            app: 应用实例
+            config: 应用配置
+        """
+        pass
+
+    def uninstall(self, app: FoundationApp) -> None:
+        """卸载插件（可选，在 lifespan 关闭时调用）。
+
+        用于清理资源。默认不做任何操作。
+
+        Args:
+            app: 应用实例
+        """
+        pass
+
+
 class Component(ABC):
     """基础设施组件基类。
 
@@ -195,6 +267,9 @@ class FoundationApp(FastAPI):
     # 默认中间件列表（子类可以覆盖）
     middlewares: ClassVar[list[type[Middleware] | Middleware]] = []
 
+    # 默认插件列表（子类可以覆盖）
+    plugins: ClassVar[list[type[Plugin] | Plugin]] = []
+
     # 默认组件列表（子类可以覆盖）
     components: ClassVar[list[type[Component] | Component]] = []
 
@@ -249,8 +324,9 @@ class FoundationApp(FastAPI):
         # 注册 access 日志（HTTP 请求日志）
         register_log_sink("access", filter_key="access")
 
-        # 初始化中间件和组件管理
+        # 初始化中间件、插件和组件管理
         self._middlewares: dict[str, Middleware] = {}
+        self._plugins: dict[str, Plugin] = {}
         self._components: dict[str, Component] = {}
         self._lifecycle_listeners: dict[str, list[Callable]] = {}
 
@@ -263,9 +339,12 @@ class FoundationApp(FastAPI):
             yield
             # 关闭
             await self._on_shutdown()
+            # 卸载插件
+            self._uninstall_plugins()
 
-        # 收集中间件和组件实例并过滤
+        # 收集中间件、插件和组件实例并过滤
         self._collect_middlewares()
+        self._collect_plugins()
         self._collect_components()
 
         # 构建中间件实例列表，传给 FastAPI
@@ -281,6 +360,9 @@ class FoundationApp(FastAPI):
             **kwargs,
         )
 
+        # 安装插件（app 创建后、lifespan 之前）
+        self._install_plugins()
+
         # 异常处理：显式注册以覆盖 FastAPI/Starlette 默认处理器，确保统一响应格式
         self.add_exception_handler(RequestValidationError, global_exception_handler)  # 422 参数校验
         self.add_exception_handler(HTTPException, global_exception_handler)  # 4xx/5xx HTTP 异常
@@ -306,6 +388,22 @@ class FoundationApp(FastAPI):
                 self._middlewares[middleware.name] = middleware
                 logger.debug(f"中间件已收集: {middleware.name}")
 
+    def _collect_plugins(self) -> None:
+        """收集并实例化所有插件。
+
+        这一步在 super().__init__() 之前执行，只做实例化和过滤。
+        """
+        for item in self.plugins:
+            # 支持类或实例
+            if isinstance(item, type):
+                plugin = item()
+            else:
+                plugin = item
+
+            if plugin.can_enable(self._config):
+                self._plugins[plugin.name] = plugin
+                logger.debug(f"插件已收集: {plugin.name}")
+
     def _collect_components(self) -> None:
         """收集并实例化所有组件。
 
@@ -323,6 +421,32 @@ class FoundationApp(FastAPI):
                 self._components[component.name] = component
                 logger.debug(f"组件已收集: {component.name}")
 
+    def _install_plugins(self) -> None:
+        """安装所有插件。
+
+        在 super().__init__() 之后、lifespan 之前同步调用。
+        此时 app 已创建，插件可以访问 app 实例。
+        """
+        for plugin in self._plugins.values():
+            try:
+                plugin.install(self, self._config)
+                logger.info(f"插件已安装: {plugin.name}")
+            except Exception as e:
+                logger.warning(f"插件安装失败 ({plugin.name}): {e}")
+
+    def _uninstall_plugins(self) -> None:
+        """卸载所有插件。
+
+        在 lifespan 关闭后同步调用。
+        """
+        # 反序卸载
+        for plugin in reversed(list(self._plugins.values())):
+            try:
+                plugin.uninstall(self)
+                logger.debug(f"插件已卸载: {plugin.name}")
+            except Exception as e:
+                logger.warning(f"插件卸载失败 ({plugin.name}): {e}")
+
     def _build_middlewares(self) -> list[StarletteMiddleware]:
         """构建所有中间件实例。
 

{aury_boot-0.0.29 → aury_boot-0.0.31}/aury/boot/application/app/components.py

@@ -9,7 +9,7 @@ import importlib
 import pkgutil
 from typing import ClassVar
 
-from aury.boot.application.app.base import Component, FoundationApp
+from aury.boot.application.app.base import Component, FoundationApp, Plugin
 from aury.boot.application.config import BaseConfig
 from aury.boot.application.constants import ComponentName, ServiceType
 from aury.boot.application.migrations import MigrationManager
@@ -18,6 +18,17 @@ from aury.boot.infrastructure.cache import CacheManager
 from aury.boot.infrastructure.channel import ChannelManager
 from aury.boot.infrastructure.database import DatabaseManager
 from aury.boot.infrastructure.events import EventBusManager
+from aury.boot.infrastructure.monitoring.alerting import (
+    AlertEventType,
+    AlertManager,
+    AlertSeverity,
+    emit_alert,
+)
+from aury.boot.infrastructure.monitoring.tracing import (
+    TelemetryConfig,
+    TelemetryProvider,
+    setup_otel_logging,
+)
 from aury.boot.infrastructure.mq import MQManager
 from aury.boot.infrastructure.scheduler import SchedulerManager
 from aury.boot.infrastructure.storage import StorageManager
@@ -48,9 +59,26 @@ class DatabaseComponent(Component):
                     pool_timeout=config.database.pool_timeout,
                     pool_recycle=config.database.pool_recycle,
                 )
+                # 在 engine 创建后进行 OTEL instrumentation
+                self._instrument_engine(db_manager, config)
         except Exception as e:
             logger.error(f"数据库初始化失败: {e}")
             raise
+    
+    def _instrument_engine(self, db_manager: DatabaseManager, config: BaseConfig) -> None:
+        """对数据库引擎进行 OTEL instrumentation。"""
+        if not config.telemetry.enabled:
+            return
+        
+        try:
+            from opentelemetry.instrumentation.sqlalchemy import SQLAlchemyInstrumentor
+            # async engine 需要使用 sync_engine
+            SQLAlchemyInstrumentor().instrument(engine=db_manager.engine.sync_engine)
+            logger.debug("SQLAlchemy engine instrumentation 已启用")
+        except ImportError:
+            logger.debug("SQLAlchemy instrumentation 未安装，跳过")
+        except Exception as e:
+            logger.warning(f"SQLAlchemy engine instrumentation 失败: {e}")
 
     async def teardown(self, app: FoundationApp) -> None:
         """关闭数据库。"""
@@ -549,6 +577,195 @@ class ChannelComponent(Component):
                 logger.warning(f"通道 [{name}] 关闭失败: {e}")
 
 
+class TelemetryPlugin(Plugin):
+    """OpenTelemetry 插件。
+    
+    在 app 创建后同步执行，负责：
+    - 初始化 TracerProvider
+    - 对 FastAPI app 进行 instrumentation
+    - 配置 AlertingSpanProcessor 检测慢请求/异常（告警配置从 ALERT__ 读取）
+    
+    配置方式（环境变量）：
+        TELEMETRY__ENABLED=true
+        TELEMETRY__TRACES_ENDPOINT=http://jaeger:4317  # 可选
+        TELEMETRY__LOGS_ENDPOINT=http://loki:3100      # 可选
+        TELEMETRY__METRICS_ENDPOINT=http://prometheus:9090  # 可选
+        
+        # 告警相关配置从 ALERT__ 读取
+        ALERT__ENABLED=true
+        ALERT__SLOW_REQUEST_THRESHOLD=1.0
+        ALERT__SLOW_SQL_THRESHOLD=0.5
+        ALERT__ALERT_ON_SLOW_REQUEST=true
+        ALERT__ALERT_ON_SLOW_SQL=true
+        ALERT__ALERT_ON_ERROR=true
+    """
+
+    name = "telemetry"
+    enabled = True
+
+    def can_enable(self, config: BaseConfig) -> bool:
+        """仅当 TELEMETRY__ENABLED=true 时启用。"""
+        return self.enabled and config.telemetry.enabled
+
+    def install(self, app: FoundationApp, config: BaseConfig) -> None:
+        """初始化 OpenTelemetry 并对 app 进行 instrumentation。"""
+        try:
+            # 创建告警回调（调用 alerting 模块，自动应用定制化规则）
+            alert_callback = self._create_alert_callback() if config.alert.enabled else None
+            
+            telemetry_config = TelemetryConfig(
+                service_name=config.service.name,
+                service_version=config.service.version,
+                environment=config.service.environment,
+                instrument_fastapi=config.telemetry.instrument_fastapi,
+                instrument_sqlalchemy=config.telemetry.instrument_sqlalchemy,
+                instrument_httpx=config.telemetry.instrument_httpx,
+                # 告警配置全部从 AlertSettings 读取
+                alert_enabled=config.alert.enabled,
+                slow_request_threshold=config.alert.slow_request_threshold,
+                slow_sql_threshold=config.alert.slow_sql_threshold,
+                alert_on_slow_request=config.alert.alert_on_slow_request,
+                alert_on_slow_sql=config.alert.alert_on_slow_sql,
+                alert_on_error=config.alert.alert_on_error,
+                alert_callback=alert_callback,
+                slow_request_exclude_paths=config.alert.slow_request_exclude_paths,
+                traces_endpoint=config.telemetry.traces_endpoint,
+                traces_headers=config.telemetry.traces_headers,
+                logs_endpoint=config.telemetry.logs_endpoint,
+                logs_headers=config.telemetry.logs_headers,
+                metrics_endpoint=config.telemetry.metrics_endpoint,
+                metrics_headers=config.telemetry.metrics_headers,
+                sampling_rate=config.telemetry.sampling_rate,
+            )
+            
+            # 创建并初始化 TelemetryProvider
+            provider = TelemetryProvider(telemetry_config)
+            success = provider.initialize()
+            
+            if success:
+                # 保存到 app.state 以便 shutdown 时使用
+                app.state.telemetry_provider = provider
+                
+                # 对已创建的 FastAPI app 进行 instrumentation
+                provider.instrument_fastapi_app(app)
+                
+                # 配置 OTLP 日志导出（可选）
+                if config.telemetry.logs_endpoint:
+                    self._setup_otlp_logging(
+                        config.telemetry.logs_endpoint,
+                        config.telemetry.logs_headers,
+                    )
+            else:
+                logger.warning("OpenTelemetry 初始化失败（非关键）")
+                
+        except ImportError as e:
+            logger.warning(f"OpenTelemetry 未安装，跳过初始化: {e}")
+        except Exception as e:
+            logger.warning(f"OpenTelemetry 初始化失败（非关键）: {e}")
+
+    def _create_alert_callback(self):
+        """创建告警回调函数。
+        
+        回调会调用 alerting 模块的 emit_alert，自动应用 YAML 定制化规则。
+        """
+        # 事件类型映射
+        type_mapping = {
+            "slow_request": AlertEventType.SLOW_REQUEST,
+            "slow_sql": AlertEventType.SLOW_SQL,
+            "exception": AlertEventType.EXCEPTION,
+            "custom": AlertEventType.CUSTOM,
+        }
+        
+        # 严重级别映射
+        severity_mapping = {
+            "info": AlertSeverity.INFO,
+            "warning": AlertSeverity.WARNING,
+            "error": AlertSeverity.ERROR,
+            "critical": AlertSeverity.CRITICAL,
+        }
+        
+        async def callback(event_type: str, message: str, **metadata):
+            try:
+                alert_event_type = type_mapping.get(event_type, AlertEventType.CUSTOM)
+                severity_str = metadata.pop("severity", "warning")
+                severity = severity_mapping.get(severity_str, AlertSeverity.WARNING)
+                
+                await emit_alert(
+                    alert_event_type,
+                    message,
+                    severity=severity,
+                    **metadata,
+                )
+            except Exception as e:
+                logger.debug(f"发送告警失败: {e}")
+        
+        return callback
+    
+    def _setup_otlp_logging(
+        self,
+        endpoint: str,
+        headers: dict[str, str] | None = None,
+    ) -> None:
+        """配置 OTLP 日志导出。"""
+        try:
+            setup_otel_logging(endpoint=endpoint, headers=headers)
+        except Exception as e:
+            logger.warning(f"OTLP 日志导出配置失败: {e}")
+
+    def uninstall(self, app: FoundationApp) -> None:
+        """关闭 OpenTelemetry。"""
+        try:
+            provider = getattr(app.state, "telemetry_provider", None)
+            if provider:
+                provider.shutdown()
+                logger.info("OpenTelemetry 已关闭")
+        except Exception as e:
+            logger.warning(f"OpenTelemetry 关闭失败: {e}")
+
+
+class AlertComponent(Component):
+    """告警系统组件。
+    
+    在 lifespan 启动时异步初始化 AlertManager。
+    """
+
+    name = "alert"
+    enabled = True
+    depends_on: ClassVar[list[str]] = []
+
+    def can_enable(self, config: BaseConfig) -> bool:
+        """仅当 ALERT__ENABLED=true 时启用。"""
+        return self.enabled and config.alert.enabled
+
+    async def setup(self, app: FoundationApp, config: BaseConfig) -> None:
+        """初始化告警管理器。"""
+        try:
+            alert_manager = AlertManager.get_instance()
+            if not alert_manager.is_initialized:
+                await alert_manager.initialize(
+                    enabled=config.alert.enabled,
+                    service_name=config.service.name,
+                    rules_file=config.alert.rules_file,
+                    defaults={
+                        "slow_request_threshold": config.alert.slow_request_threshold,
+                        "slow_sql_threshold": config.alert.slow_sql_threshold,
+                        "aggregate_window": config.alert.aggregate_window,
+                        "slow_request_aggregate": config.alert.slow_request_aggregate,
+                        "slow_sql_aggregate": config.alert.slow_sql_aggregate,
+                        "exception_aggregate": config.alert.exception_aggregate,
+                        "suppress_seconds": config.alert.suppress_seconds,
+                        "slow_request_exclude_paths": config.alert.slow_request_exclude_paths,
+                    },
+                    notifiers=config.alert.get_notifiers(),
+                )
+        except Exception as e:
+            logger.warning(f"告警管理器初始化失败（非关键）: {e}")
+
+    async def teardown(self, app: FoundationApp) -> None:
+        """无需清理。"""
+        pass
+
+
 class EventBusComponent(Component):
     """事件总线组件。
     
@@ -590,8 +807,14 @@ class EventBusComponent(Component):
                 logger.warning(f"事件总线 [{name}] 关闭失败: {e}")
 
 
+# 设置默认插件
+FoundationApp.plugins = [
+    TelemetryPlugin,  # app 创建后立即 instrument
+]
+
 # 设置默认组件
 FoundationApp.components = [
+    AlertComponent,  # 最先初始化告警管理器
     DatabaseComponent,
     MigrationComponent,
     AdminConsoleComponent,
@@ -607,6 +830,7 @@ FoundationApp.components = [
 
 __all__ = [
     "AdminConsoleComponent",
+    "AlertComponent",
     "CacheComponent",
     "ChannelComponent",
     "DatabaseComponent",
@@ -616,5 +840,6 @@ __all__ = [
     "SchedulerComponent",
     "StorageComponent",
     "TaskComponent",
+    "TelemetryPlugin",
 ]