数据库监控日志、接口监控，熔断机制，提高连接池

app/core/database.py

@@ -1,14 +1,28 @@
 from sqlalchemy import create_engine, MetaData
 from sqlalchemy.ext.declarative import declarative_base
 from sqlalchemy.orm import sessionmaker
+from sqlalchemy.pool import QueuePool
 from .config import settings
 
-engine = create_engine(settings.DATABASE_URL, pool_pre_ping=True, echo=False, pool_size=30, max_overflow=60, pool_timeout=30, pool_recycle=3600)
+# 创建带连接池监控的引擎
+engine = create_engine(
+    settings.DATABASE_URL,
+    poolclass=QueuePool,
+    pool_pre_ping=True,
+    echo=False,  # 生产环境建议关闭SQL日志
+    pool_size=400,
+    max_overflow=600,
+    pool_timeout=60,  # 增加超时时间到60秒
+    pool_recycle=3600,  # 1小时回收连接
+    pool_reset_on_return='commit'  # 归还连接时重置状态
+)
+
 SessionLocal = sessionmaker(autocommit=False, autoflush=False, bind=engine)
 
 Base = declarative_base()
 
 def get_db():
+    """数据库依赖注入函数"""
     db = SessionLocal()
     try:
         yield db

app/core/db_monitor.py

@@ -0,0 +1,252 @@
+"""
+数据库连接池监控模块
+监控连接池状态、事务执行情况，预防雪崩效应
+"""
+import logging
+import time
+import threading
+from sqlalchemy import create_engine, event
+from sqlalchemy.engine import Engine
+from sqlalchemy.pool import QueuePool
+from typing import Dict, Any, Optional
+from datetime import datetime, timedelta
+from .database import engine
+from .logging_config import get_logger
+
+logger = get_logger(__name__)
+
+# 全局监控数据
+_pool_stats = {
+    'total_connections': 0,
+    'checked_in': 0,
+    'checked_out': 0,
+    'overflow': 0,
+    'invalidate_count': 0,
+    'transactions': [],  # 事务统计
+    'slow_queries': [],  # 慢查询
+    'connection_errors': [],  # 连接错误
+    'peak_connections': 0,
+    'last_reset': datetime.now()
+}
+
+# 告警配置
+_alert_thresholds = {
+    'pool_usage_percent': 80,  # 连接池使用率告警阈值
+    'slow_query_time': 5.0,  # 慢查询阈值（秒）
+    'max_transaction_time': 30.0,  # 最大事务执行时间
+    'connection_error_count': 10,  # 连接错误告警阈值
+    'alert_cooldown': 300  # 告警冷却时间（秒）
+}
+
+_last_alerts = {}
+
+def get_pool_status() -> Dict[str, Any]:
+    """获取连接池状态"""
+    if hasattr(engine.pool, 'status'):
+        pool = engine.pool
+        stats = {
+            'total': pool.size() if hasattr(pool, 'size') else 0,
+            'checked_in': pool.checkedin() if hasattr(pool, 'checkedin') else 0,
+            'checked_out': pool.checkedout() if hasattr(pool, 'checkedout') else 0,
+            'overflow': pool.overflow() if hasattr(pool, 'overflow') else 0,
+            'invalidate': pool.invalid() if hasattr(pool, 'invalid') else 0,
+        }
+    else:
+        # 估算值
+        stats = {
+            'total': _pool_stats.get('total_connections', 0),
+            'checked_in': _pool_stats.get('checked_in', 0),
+            'checked_out': _pool_stats.get('checked_out', 0),
+            'overflow': _pool_stats.get('overflow', 0),
+            'invalidate': _pool_stats.get('invalidate_count', 0),
+        }
+
+    # 计算使用率
+    if stats['total'] > 0:
+        usage_percent = (stats['checked_out'] / stats['total']) * 100
+        stats['usage_percent'] = round(usage_percent, 2)
+    else:
+        stats['usage_percent'] = 0
+
+    # 更新峰值
+    if stats['checked_out'] > _pool_stats['peak_connections']:
+        _pool_stats['peak_connections'] = stats['checked_out']
+
+    return stats
+
+def check_pool_alerts():
+    """检查连接池告警"""
+    current_time = time.time()
+    stats = get_pool_status()
+
+    # 连接池使用率告警
+    if stats.get('usage_percent', 0) >= _alert_thresholds['pool_usage_percent']:
+        alert_key = 'pool_usage'
+        if alert_key not in _last_alerts or (current_time - _last_alerts.get(alert_key, 0)) > _alert_thresholds['alert_cooldown']:
+            logger.warning(
+                f"🚨 数据库连接池告警: 使用率 {stats['usage_percent']}% 超过阈值 {_alert_thresholds['pool_usage_percent']}% "
+                f"(已使用: {stats['checked_out']}/{stats['total']})"
+            )
+            _last_alerts[alert_key] = current_time
+
+    # 连接错误告警
+    error_count = len(_pool_stats['connection_errors'])
+    if error_count >= _alert_thresholds['connection_error_count']:
+        alert_key = 'connection_errors'
+        if alert_key not in _last_alerts or (current_time - _last_alerts.get(alert_key, 0)) > _alert_thresholds['alert_cooldown']:
+            recent_errors = [e for e in _pool_stats['connection_errors'] if (current_time - e['timestamp']) < 300]
+            logger.warning(
+                f"🚨 数据库连接错误告警: 近5分钟内发生 {len(recent_errors)} 次连接错误"
+            )
+            _last_alerts[alert_key] = current_time
+
+    # 慢查询告警
+    slow_queries = [q for q in _pool_stats['slow_queries'] if (current_time - q['timestamp']) < 300]
+    if len(slow_queries) >= 5:
+        alert_key = 'slow_queries'
+        if alert_key not in _last_alerts or (current_time - _last_alerts.get(alert_key, 0)) > _alert_thresholds['alert_cooldown']:
+            avg_time = sum(q['duration'] for q in slow_queries) / len(slow_queries)
+            logger.warning(
+                f"🚨 慢查询告警: 近5分钟内 {len(slow_queries)} 个慢查询，平均耗时 {avg_time:.2f}s"
+            )
+            _last_alerts[alert_key] = current_time
+
+def log_transaction_start(sql: str, params: Optional[Dict] = None):
+    """记录事务开始"""
+    transaction_info = {
+        'sql': sql[:100] + '...' if len(sql) > 100 else sql,
+        'params': params,
+        'start_time': time.time(),
+        'thread_id': threading.get_ident()
+    }
+    _pool_stats['transactions'].append(transaction_info)
+
+def log_transaction_end(success: bool = True, error: Optional[str] = None):
+    """记录事务结束"""
+    if not _pool_stats['transactions']:
+        return
+
+    current_time = time.time()
+    transaction = _pool_stats['transactions'][-1]
+    duration = current_time - transaction['start_time']
+
+    # 慢事务告警
+    if duration >= _alert_thresholds['max_transaction_time']:
+        logger.warning(
+            f"🐌 慢事务告警: 执行时间 {duration:.2f}s 超过阈值 {_alert_thresholds['max_transaction_time']}s "
+            f"SQL: {transaction['sql']}"
+        )
+
+    # 记录慢查询
+    if duration >= _alert_thresholds['slow_query_time']:
+        _pool_stats['slow_queries'].append({
+            'sql': transaction['sql'],
+            'duration': duration,
+            'timestamp': current_time
+        })
+
+    # 清理旧记录（保留最近1000条）
+    if len(_pool_stats['slow_queries']) > 1000:
+        _pool_stats['slow_queries'] = _pool_stats['slow_queries'][-1000:]
+
+    _pool_stats['transactions'].pop()
+
+def log_connection_error(error: str, sql: Optional[str] = None):
+    """记录连接错误"""
+    _pool_stats['connection_errors'].append({
+        'error': error,
+        'sql': sql,
+        'timestamp': time.time()
+    })
+
+    # 清理旧记录（保留最近100条）
+    if len(_pool_stats['connection_errors']) > 100:
+        _pool_stats['connection_errors'] = _pool_stats['connection_errors'][-100:]
+
+def reset_stats():
+    """重置统计信息"""
+    global _pool_stats
+    _pool_stats = {
+        'total_connections': 0,
+        'checked_in': 0,
+        'checked_out': 0,
+        'overflow': 0,
+        'invalidate_count': 0,
+        'transactions': [],
+        'slow_queries': [],
+        'connection_errors': [],
+        'peak_connections': 0,
+        'last_reset': datetime.now()
+    }
+    logger.info("数据库监控统计已重置")
+
+def get_monitoring_report() -> Dict[str, Any]:
+    """获取监控报告"""
+    stats = get_pool_status()
+    current_time = time.time()
+
+    # 计算最近5分钟的数据
+    recent_slow_queries = [q for q in _pool_stats['slow_queries'] if (current_time - q['timestamp']) < 300]
+    recent_errors = [e for e in _pool_stats['connection_errors'] if (current_time - e['timestamp']) < 300]
+
+    report = {
+        'timestamp': datetime.now().isoformat(),
+        'pool_status': stats,
+        'peak_connections': _pool_stats['peak_connections'],
+        'recent_5min': {
+            'slow_queries_count': len(recent_slow_queries),
+            'connection_errors_count': len(recent_errors),
+            'avg_slow_query_time': sum(q['duration'] for q in recent_slow_queries) / len(recent_slow_queries) if recent_slow_queries else 0
+        },
+        'slow_queries': recent_slow_queries[-10:],  # 最近10条慢查询
+        'connection_errors': recent_errors[-10:],  # 最近10条连接错误
+        'last_reset': _pool_stats['last_reset'].isoformat()
+    }
+
+    return report
+
+# 定时监控任务
+def monitoring_task():
+    """定时监控任务"""
+    while True:
+        try:
+            check_pool_alerts()
+            time.sleep(30)  # 每30秒检查一次
+        except Exception as e:
+            logger.error(f"数据库监控任务异常: {e}")
+            time.sleep(60)  # 异常时等待更长时间
+
+# 启动后台监控线程
+def start_monitoring():
+    """启动后台监控"""
+    monitor_thread = threading.Thread(target=monitoring_task, daemon=True)
+    monitor_thread.start()
+    logger.info("数据库连接池监控已启动")
+
+# SQLAlchemy事件监听
+@event.listens_for(Engine, "before_cursor_execute")
+def receive_before_cursor_execute(conn, cursor, statement, params, context, executemany):
+    """SQL执行前监听"""
+    log_transaction_start(statement, params)
+
+@event.listens_for(Engine, "after_cursor_execute")
+def receive_after_cursor_execute(conn, cursor, statement, params, context, executemany):
+    """SQL执行后监听"""
+    log_transaction_end(success=True)
+
+@event.listens_for(Engine, "handle_error")
+def receive_handle_error(exception, context):
+    """错误监听"""
+    error_msg = str(exception)
+    sql = context.statement if context and hasattr(context, 'statement') else None
+    log_connection_error(error_msg, sql)
+    log_transaction_end(success=False, error=error_msg)
+
+def log_pool_status():
+    """记录连接池状态到日志"""
+    stats = get_pool_status()
+    logger.info(
+        f"数据库连接池状态: 使用率 {stats['usage_percent']}% "
+        f"(已用: {stats['checked_out']}, 空闲: {stats['checked_in']}, 总计: {stats['total']}) "
+        f"峰值: {_pool_stats['peak_connections']}"
+    )

app/core/retry.py

@@ -0,0 +1,211 @@
+"""
+重试机制和雪崩效应防护
+提供指数退避、熔断器、重试装饰器等功能
+"""
+import logging
+import time
+import functools
+from typing import Callable, Any, Optional
+from enum import Enum
+from datetime import datetime, timedelta
+
+logger = logging.getLogger(__name__)
+
+class CircuitBreakerState(Enum):
+    """熔断器状态"""
+    CLOSED = "closed"      # 正常状态
+    OPEN = "open"          # 熔断状态
+    HALF_OPEN = "half_open"  # 半开状态
+
+class CircuitBreaker:
+    """熔断器实现"""
+
+    def __init__(self, failure_threshold: int = 5, recovery_timeout: int = 60):
+        """
+        初始化熔断器
+
+        Args:
+            failure_threshold: 失败阈值，达到此数量后触发熔断
+            recovery_timeout: 恢复超时时间（秒）
+        """
+        self.failure_threshold = failure_threshold
+        self.recovery_timeout = recovery_timeout
+        self.failure_count = 0
+        self.last_failure_time = None
+        self.state = CircuitBreakerState.CLOSED
+
+    def call(self, func: Callable, *args, **kwargs):
+        """通过熔断器执行函数"""
+        if self.state == CircuitBreakerState.OPEN:
+            # 检查是否应该进入半开状态
+            if self.last_failure_time and \
+               (datetime.now() - self.last_failure_time).seconds >= self.recovery_timeout:
+                self.state = CircuitBreakerState.HALF_OPEN
+                logger.info("熔断器进入半开状态")
+            else:
+                raise Exception("熔断器开启，直接拒绝请求")
+
+        try:
+            result = func(*args, **kwargs)
+            # 执行成功，重置状态
+            if self.state == CircuitBreakerState.HALF_OPEN:
+                self.state = CircuitBreakerState.CLOSED
+                self.failure_count = 0
+                logger.info("熔断器关闭，恢复正常")
+            return result
+        except Exception as e:
+            self.failure_count += 1
+            self.last_failure_time = datetime.now()
+
+            if self.failure_count >= self.failure_threshold:
+                self.state = CircuitBreakerState.OPEN
+                logger.warning(f"熔断器开启，失败次数: {self.failure_count}")
+
+            raise e
+
+class RetryConfig:
+    """重试配置"""
+    def __init__(
+        self,
+        max_attempts: int = 3,
+        base_delay: float = 1.0,
+        max_delay: float = 60.0,
+        exponential_base: float = 2.0,
+        jitter: bool = True
+    ):
+        """
+        初始化重试配置
+
+        Args:
+            max_attempts: 最大重试次数
+            base_delay: 基础延迟时间（秒）
+            max_delay: 最大延迟时间（秒）
+            exponential_base: 指数退避基数
+            jitter: 是否添加随机抖动
+        """
+        self.max_attempts = max_attempts
+        self.base_delay = base_delay
+        self.max_delay = max_delay
+        self.exponential_base = exponential_base
+        self.jitter = jitter
+
+def retry(
+    config: Optional[RetryConfig] = None,
+    exceptions: tuple = (Exception,)
+):
+    """
+    重试装饰器
+
+    Args:
+        config: 重试配置，如果为None则使用默认配置
+        exceptions: 需要重试的异常类型
+    """
+    if config is None:
+        config = RetryConfig()
+
+    def decorator(func: Callable) -> Callable:
+        @functools.wraps(func)
+        def wrapper(*args, **kwargs):
+            last_exception = None
+
+            for attempt in range(config.max_attempts):
+                try:
+                    return func(*args, **kwargs)
+                except exceptions as e:
+                    last_exception = e
+
+                    if attempt == config.max_attempts - 1:
+                        # 最后一次尝试失败，抛出异常
+                        logger.error(
+                            f"函数 {func.__name__} 经过 {config.max_attempts} 次重试后仍然失败: {str(e)}"
+                        )
+                        raise e
+
+                    # 计算延迟时间
+                    delay = min(
+                        config.base_delay * (config.exponential_base ** attempt),
+                        config.max_delay
+                    )
+
+                    # 添加抖动
+                    if config.jitter:
+                        import random
+                        delay = delay * (0.5 + random.random() * 0.5)
+
+                    logger.warning(
+                        f"函数 {func.__name__} 第 {attempt + 1} 次尝试失败: {str(e)}, "
+                        f"{delay:.2f} 秒后重试"
+                    )
+
+                    time.sleep(delay)
+
+            # 理论上不会到达这里，但为了安全起见
+            if last_exception:
+                raise last_exception
+
+        return wrapper
+    return decorator
+
+# 全局熔断器实例
+_default_circuit_breaker = CircuitBreaker()
+
+def circuit_breaker(
+    failure_threshold: int = 5,
+    recovery_timeout: int = 60
+):
+    """
+    熔断器装饰器
+
+    Args:
+        failure_threshold: 失败阈值
+        recovery_timeout: 恢复超时时间
+    """
+    def decorator(func: Callable) -> Callable:
+        breaker = CircuitBreaker(failure_threshold, recovery_timeout)
+
+        @functools.wraps(func)
+        def wrapper(*args, **kwargs):
+            return breaker.call(func, *args, **kwargs)
+
+        # 将熔断器实例附加到函数上，方便外部查看状态
+        wrapper.circuit_breaker = breaker
+        return wrapper
+
+    return decorator
+
+def with_circuit_breaker(func: Callable, *args, **kwargs):
+    """使用默认熔断器执行函数"""
+    return _default_circuit_breaker.call(func, *args, **kwargs)
+
+# 预定义的重试配置
+RETRY_CONFIG_FAST = RetryConfig(
+    max_attempts=3,
+    base_delay=0.5,
+    max_delay=5.0,
+    exponential_base=2.0,
+    jitter=True
+)
+
+RETRY_CONFIG_SLOW = RetryConfig(
+    max_attempts=5,
+    base_delay=2.0,
+    max_delay=60.0,
+    exponential_base=2.0,
+    jitter=True
+)
+
+RETRY_CONFIG_DB = RetryConfig(
+    max_attempts=3,
+    base_delay=1.0,
+    max_delay=10.0,
+    exponential_base=2.0,
+    jitter=True
+)
+
+# 数据库操作重试装饰器
+def retry_db_operation(max_attempts: int = 3):
+    """数据库操作重试装饰器"""
+    return retry(
+        config=RETRY_CONFIG_DB,
+        exceptions=(Exception,)
+    )

app/main.py

@@ -6,6 +6,7 @@ import logging
 from .core.config import settings
 from .core.logging_config import setup_logging, get_logger
 from .core.database import init_db
+from .core.db_monitor import start_monitoring, log_pool_status, get_pool_status
 from .api.account import router as account_router
 from .api.database import router as database_router
 from .api.task import router as task_router
@@ -34,6 +35,17 @@ async def lifespan(app: FastAPI):
     except Exception as e:
         logger.error(f"数据库初始化失败: {e}")
 
+    # 启动数据库监控
+    try:
+        start_monitoring()
+        logger.info("数据库连接池监控已启动")
+
+        # 记录初始连接池状态
+        pool_stats = get_pool_status()
+        logger.info(f"初始连接池状态: {pool_stats}")
+    except Exception as e:
+        logger.error(f"数据库监控启动失败: {e}")
+
     # 启动定时任务调度器
     try:
         task_scheduler.start()
@@ -45,6 +57,14 @@ async def lifespan(app: FastAPI):
 
     # 关闭时执行
     logger.info("应用关闭中...")
+
+    # 记录最终连接池状态
+    try:
+        pool_stats = get_pool_status()
+        logger.info(f"最终连接池状态: {pool_stats}")
+    except Exception as e:
+        logger.error(f"获取最终连接池状态失败: {e}")
+
     try:
         task_scheduler.shutdown()
         logger.info("定时任务调度器已关闭")
@@ -101,11 +121,47 @@ async def health_check():
         "scheduler": "running" if task_scheduler.scheduler.running else "stopped"
     }
 
+# 数据库监控端点
+@app.get("/api/monitor/database")
+async def get_database_monitor():
+    """获取数据库连接池监控信息"""
+    from .core.db_monitor import get_monitoring_report
+    return get_monitoring_report()
+
+# 连接池状态端点
+@app.get("/api/monitor/pool")
+async def get_pool_status():
+    """获取连接池状态"""
+    from .core.db_monitor import get_pool_status
+    return get_pool_status()
+
 # 全局异常处理
 @app.exception_handler(Exception)
 async def global_exception_handler(request, exc):
-    logger.error(f"全局异常: {str(exc)}")
+    """全局异常处理"""
+    import traceback
+    from .core.db_monitor import get_pool_status
+
+    # 获取异常详情
+    error_msg = str(exc)
+    error_type = type(exc).__name__
+    stack_trace = traceback.format_exc()
+
+    # 获取当前连接池状态
+    try:
+        pool_stats = get_pool_status()
+        pool_info = f", 连接池使用率: {pool_stats.get('usage_percent', 'N/A')}%"
+    except:
+        pool_info = ""
+
+    # 记录详细错误日志
+    logger.error(
+        f"🚨 全局异常: {error_type}: {error_msg} "
+        f"请求路径: {request.url.path}, 方法: {request.method}{pool_info}\n"
+        f"堆栈跟踪:\n{stack_trace}"
+    )
+
     return HTTPException(
         status_code=500,
-        detail="服务器内部错误"
+        detail=f"服务器内部错误: {error_type}"
     )

app/services/original_data.py

@@ -1,11 +1,14 @@
 from sqlalchemy.orm import Session
 from sqlalchemy import text, inspect
+from sqlalchemy.exc import SQLAlchemyError, DisconnectionError, TimeoutError
 from typing import List, Optional, Dict, Any
 from ..models.original_data import OriginalData, get_original_data_model, get_table_name
 from .base import BaseService
 from ..models.settlement_data import SettlementData
 from ..models.account import Account
 from ..core.database import engine
+from ..core.db_monitor import log_pool_status, get_pool_status
+from ..core.retry import retry, circuit_breaker, RetryConfig
 import logging
 
 logger = logging.getLogger(__name__)
@@ -363,49 +366,84 @@ class OriginalDataService(BaseService[OriginalData]):
 
         table_name = self._get_table_name(account_id)
 
-        # **重要**: 始终使用内部事务，确保数据能正确提交
-        # 这是为了解决外部事务可能不提交的问题
-        # in_transaction = db.in_transaction()
+        # 记录开始前的连接池状态
+        pool_stats_before = get_pool_status()
+        logger.info(f"开始批量导入，连接池状态: {pool_stats_before}")
 
-        # 始终创建内部事务
-        for attempt in range(2):  # 最多重试1次
-            try:
-                db.begin()
-                success_count = 0
-                failed_count = 0
-                failed_items = []
+        # 检查是否已在事务中
+        in_transaction = db.in_transaction()
+        logger.info(f"当前事务状态: {'已在事务中' if in_transaction else '无事务'}")
 
-                # 执行数据导入操作
-                success_count = self._execute_import(db, table_name, data, account_id)
-                db.commit()
-                logger.info(f"Batch import original data completed. Success: {success_count}, Failed: {failed_count}")
-                break
-
-            except Exception as e:
+        @retry(config=RetryConfig(max_attempts=2, base_delay=2.0, max_delay=10.0))
+        def _do_import():
+            """执行导入操作的内部函数（带重试）"""
+            for attempt in range(2):  # 最多重试1次
                 try:
-                    db.rollback()
-                except:
-                    pass  # 如果回滚失败，忽略错误
-                logger.warning(f"Batch import attempt {attempt + 1} failed: {str(e)}")
-                if attempt == 1:  # 最后一次重试失败
-                    logger.error("Batch import original data failed after retries")
+                    # 只有不在事务中时才调用begin()
+                    if not in_transaction:
+                        logger.info(f"开始内部事务 (尝试 {attempt + 1})")
+                        db.begin()
+                    else:
+                        logger.info(f"使用外部事务执行导入 (尝试 {attempt + 1})")
+
+                    success_count = 0
+                    failed_count = 0
+                    failed_items = []
+
+                    # 执行数据导入操作
+                    success_count = self._execute_import(db, table_name, data, account_id)
+
+                    # 只有我们开始的事务才提交
+                    if not in_transaction:
+                        db.commit()
+                        logger.info(f"事务已提交")
+                    else:
+                        logger.info(f"使用外部事务，不提交")
+
+                    logger.info(f"Batch import original data completed. Success: {success_count}, Failed: {failed_count}")
                     return {
-                        'success': False,
-                        'message': f'批量导入失败: {str(e)}',
+                        'success': True,
+                        'message': '批量导入完成' if failed_count == 0 else f'部分导入失败',
                         'total_count': total_count,
-                        'success_count': 0,
-                        'failed_count': total_count,
+                        'success_count': success_count,
+                        'failed_count': failed_count,
                         'failed_items': failed_items
                     }
 
-        return {
-            'success': True,
-            'message': '批量导入完成' if failed_count == 0 else f'部分导入失败',
-            'total_count': total_count,
-            'success_count': success_count,
-            'failed_count': failed_count,
-            'failed_items': failed_items
-        }
+                except SQLAlchemyError as e:
+                    # 只有我们开始的事务才回滚
+                    if not in_transaction:
+                        try:
+                            db.rollback()
+                        except:
+                            pass
+
+                    pool_stats_after = get_pool_status()
+                    error_msg = f"数据库错误 (尝试 {attempt + 1}): {str(e)}"
+                    logger.error(f"{error_msg}, 连接池状态: {pool_stats_after}")
+
+                    # 记录错误详情
+                    logger.error(
+                        f"错误详情: 类型={type(e).__name__}, "
+                        f"连接池使用率={pool_stats_after.get('usage_percent', 0)}%, "
+                        f"SQL: {str(e)[:200]}"
+                    )
+
+                    if attempt == 1:  # 最后一次重试失败
+                        logger.error("Batch import original data failed after retries")
+                        raise e  # 抛出异常触发重试装饰器
+
+        try:
+            return _do_import()
+        except Exception as e:
+            return {
+                'success': False,
+                'message': f'批量导入失败: {str(e)}',
+                'total_count': total_count,
+                'success_count': 0,
+                'failed_count': total_count,
+                'failed_items': []
+            }
 
     def _execute_import(self, db: Session, table_name: str, data: List, account_id: int) -> int:
         """执行数据导入操作（抽取的公共逻辑）"""
@@ -610,16 +648,23 @@ class OriginalDataService(BaseService[OriginalData]):
         total_count = sum(len(group['data']) for group in group_validation_results if group['valid'])
         logger.info(f"Total valid groups: {len(group_validation_results)}, Total records: {total_count}")
 
-        # **重要**: 始终使用内部事务，确保数据能正确提交
-        # 这是为了解决外部事务可能不提交的问题
-        # in_transaction = db.in_transaction()
-        # logger.info(f"Original transaction status: {'in_transaction' if in_transaction else 'not in_transaction'}")
+        # 记录开始前的连接池状态
+        pool_stats_before = get_pool_status()
+        logger.info(f"开始新版批量导入，连接池状态: {pool_stats_before}")
+
+        # 检查是否已在事务中
+        in_transaction = db.in_transaction()
+        logger.info(f"当前事务状态: {'已在事务中' if in_transaction else '无事务'}")
 
-        # 始终创建内部事务
         for attempt in range(2):
             try:
-                logger.info(f"Starting internal transaction (attempt {attempt + 1})")
-                db.begin()
+                # 只有不在事务中时才调用begin()
+                if not in_transaction:
+                    logger.info(f"开始内部事务 (尝试 {attempt + 1})")
+                    db.begin()
+                else:
+                    logger.info(f"使用外部事务执行导入 (尝试 {attempt + 1})")
+
                 success_count = 0
                 failed_count = 0
                 failed_items = []
@@ -670,17 +715,36 @@ class OriginalDataService(BaseService[OriginalData]):
                     logger.info(f"Account {account_id} completed: Success={group_results['success_count']}, Failed={group_results['failed_count']}")
 
                 logger.info(f"Before commit: Success={success_count}, Failed={failed_count}")
-                db.commit()
-                logger.info(f"Transaction committed successfully! Success: {success_count}, Failed: {failed_count}")
+
+                # 只有我们开始的事务才提交
+                if not in_transaction:
+                    db.commit()
+                    logger.info(f"事务已提交")
+                else:
+                    logger.info(f"使用外部事务，不提交")
+
+                logger.info(f"Transaction completed successfully! Success: {success_count}, Failed: {failed_count}")
                 break
 
-            except Exception as e:
-                logger.error(f"Transaction rollback due to: {str(e)}")
-                try:
-                    db.rollback()
-                except:
-                    pass  # 如果回滚失败，忽略错误
-                logger.warning(f"Batch import attempt {attempt + 1} failed: {str(e)}")
+            except SQLAlchemyError as e:
+                # 只有我们开始的事务才回滚
+                if not in_transaction:
+                    try:
+                        db.rollback()
+                    except:
+                        pass
+
+                pool_stats_after = get_pool_status()
+                error_msg = f"数据库错误 (尝试 {attempt + 1}): {str(e)}"
+                logger.error(f"{error_msg}, 连接池状态: {pool_stats_after}")
+
+                # 记录错误详情
+                logger.error(
+                    f"错误详情: 类型={type(e).__name__}, "
+                    f"连接池使用率={pool_stats_after.get('usage_percent', 0)}%, "
+                    f"SQL: {str(e)[:200]}"
+                )
+
                 if attempt == 1:
                     logger.error("Batch import original data failed after retries")
                     return {