feat: 累积功能变更 — 聊天集成、租户管理、小程序更新、ETL 增强、迁移脚本

包含多个会话的累积代码变更：
- backend: AI 聊天服务、触发器调度、认证增强、WebSocket、调度器最小间隔
- admin-web: ETL 状态页、任务管理、调度配置、登录优化
- miniprogram: 看板页面、聊天集成、UI 组件、导航更新
- etl: DWS 新任务（finance_area_daily/board_cache）、连接器增强
- tenant-admin: 项目初始化
- db: 19 个迁移脚本（etl_feiqiu 11 + zqyy_app 8）
- packages/shared: 枚举和工具函数更新
- tools: 数据库工具、报表生成、健康检查
- docs: PRD/架构/部署/合约文档更新

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

apps/backend/app/ai/circuit_breaker.py

@@ -0,0 +1,116 @@
+"""熔断器 — 按 app_id 独立的断路保护。
+
+状态机：CLOSED → OPEN（连续失败达阈值）→ HALF_OPEN（超时后探测）→ CLOSED/OPEN。
+内存实现，单实例部署，不依赖外部存储。
+"""
+
+from __future__ import annotations
+
+import enum
+import time
+from dataclasses import dataclass, field
+
+
+class CircuitState(enum.Enum):
+    """熔断器状态。"""
+
+    CLOSED = "closed"  # 正常放行
+    OPEN = "open"  # 熔断中，拒绝请求
+    HALF_OPEN = "half_open"  # 探测中，放行单个请求
+
+
+@dataclass
+class _BreakerState:
+    """单个 app_id 的熔断内部状态。"""
+
+    state: CircuitState = CircuitState.CLOSED
+    failure_count: int = 0
+    last_failure_time: float = 0.0
+    last_state_change: float = field(default_factory=time.monotonic)
+
+
+class CircuitBreaker:
+    """按 app_id 独立的熔断器。
+
+    - check()：检查当前状态，OPEN 且超时自动转 HALF_OPEN
+    - record_success()：HALF_OPEN→CLOSED；CLOSED 重置失败计数
+    - record_failure()：连续达阈值→OPEN；HALF_OPEN 失败→重新 OPEN
+    """
+
+    def __init__(
+        self,
+        failure_threshold: int = 5,
+        recovery_timeout: int = 60,
+    ) -> None:
+        self._failure_threshold = failure_threshold
+        self._recovery_timeout = recovery_timeout
+        self._breakers: dict[str, _BreakerState] = {}
+
+    def _get_state(self, app_id: str) -> _BreakerState:
+        """获取或初始化指定 app_id 的状态。"""
+        if app_id not in self._breakers:
+            self._breakers[app_id] = _BreakerState()
+        return self._breakers[app_id]
+
+    def check(self, app_id: str) -> CircuitState:
+        """检查当前熔断状态。
+
+        - CLOSED / HALF_OPEN：允许通过，返回对应状态
+        - OPEN 且未超时：返回 OPEN（拒绝）
+        - OPEN 且已超时：自动转 HALF_OPEN，返回 HALF_OPEN（允许探测）
+        """
+        breaker = self._get_state(app_id)
+
+        if breaker.state == CircuitState.CLOSED:
+            return CircuitState.CLOSED
+
+        if breaker.state == CircuitState.HALF_OPEN:
+            return CircuitState.HALF_OPEN
+
+        # OPEN 状态：检查是否超过恢复超时
+        elapsed = time.monotonic() - breaker.last_failure_time
+        if elapsed >= self._recovery_timeout:
+            # 超时，转为 HALF_OPEN 探测
+            breaker.state = CircuitState.HALF_OPEN
+            breaker.last_state_change = time.monotonic()
+            return CircuitState.HALF_OPEN
+
+        return CircuitState.OPEN
+
+    def record_success(self, app_id: str) -> None:
+        """记录调用成功。
+
+        - HALF_OPEN→CLOSED（探测成功，恢复正常）
+        - CLOSED 下重置失败计数
+        """
+        breaker = self._get_state(app_id)
+
+        if breaker.state == CircuitState.HALF_OPEN:
+            breaker.state = CircuitState.CLOSED
+            breaker.failure_count = 0
+            breaker.last_state_change = time.monotonic()
+        elif breaker.state == CircuitState.CLOSED:
+            # CLOSED 状态下成功重置失败计数
+            breaker.failure_count = 0
+
+    def record_failure(self, app_id: str) -> None:
+        """记录调用失败。
+
+        - CLOSED：累加失败计数，达阈值→OPEN
+        - HALF_OPEN：探测失败→重新 OPEN
+        """
+        breaker = self._get_state(app_id)
+        now = time.monotonic()
+
+        if breaker.state == CircuitState.HALF_OPEN:
+            # 探测失败，重新熔断
+            breaker.state = CircuitState.OPEN
+            breaker.failure_count = self._failure_threshold
+            breaker.last_failure_time = now
+            breaker.last_state_change = now
+        elif breaker.state == CircuitState.CLOSED:
+            breaker.failure_count += 1
+            breaker.last_failure_time = now
+            if breaker.failure_count >= self._failure_threshold:
+                breaker.state = CircuitState.OPEN
+                breaker.last_state_change = now