jettask 0.2.14__py3-none-any.whl → 0.2.16__py3-none-any.whl

jettask/services/queue_service.py

@@ -0,0 +1,711 @@
+"""
+队列服务层
+处理队列相关的业务逻辑
+"""
+from typing import Optional, List, Dict, Any
+from datetime import datetime, timedelta, timezone
+from sqlalchemy import text
+from sqlalchemy.ext.asyncio import AsyncSession
+import logging
+import traceback
+
+logger = logging.getLogger(__name__)
+
+
+class QueueService:
+    """队列服务类"""
+    
+    @staticmethod
+    def get_base_queue_name(queue_name: str) -> str:
+        """
+        提取基础队列名（去除优先级后缀）
+        
+        Args:
+            queue_name: 完整队列名
+            
+        Returns:
+            基础队列名
+        """
+        if ':' in queue_name:
+            parts = queue_name.rsplit(':', 1)
+            if parts[-1].isdigit():
+                return parts[0]
+        return queue_name
+    
+    @staticmethod
+    async def get_queues_by_namespace(namespace_data_access, namespace: str) -> Dict[str, Any]:
+        """
+        获取指定命名空间的队列列表
+        
+        Args:
+            namespace_data_access: 命名空间数据访问实例
+            namespace: 命名空间
+            
+        Returns:
+            队列列表
+        """
+        queues_data = await namespace_data_access.get_queue_stats(namespace)
+        return {
+            "success": True,
+            "data": list(set([QueueService.get_base_queue_name(q['queue_name']) for q in queues_data]))
+        }
+    
+    @staticmethod
+    async def get_queue_flow_rates(data_access, query) -> Dict[str, Any]:
+        """
+        获取单个队列的流量速率（入队、开始执行、完成）
+        
+        Args:
+            data_access: 数据访问层实例
+            query: 时间范围查询对象
+            
+        Returns:
+            队列流量速率数据
+        """
+        # 处理时间范围
+        now = datetime.now(timezone.utc)
+        
+        if query.start_time and query.end_time:
+            # 使用提供的时间范围
+            start_time = query.start_time
+            end_time = query.end_time
+            logger.info(f"使用自定义时间范围: {start_time} 到 {end_time}")
+        else:
+            # 根据time_range参数计算时间范围
+            time_range_map = {
+                "15m": timedelta(minutes=15),
+                "30m": timedelta(minutes=30),
+                "1h": timedelta(hours=1),
+                "3h": timedelta(hours=3),
+                "6h": timedelta(hours=6),
+                "12h": timedelta(hours=12),
+                "24h": timedelta(hours=24),
+                "7d": timedelta(days=7),
+                "30d": timedelta(days=30),
+            }
+            
+            # 优先使用 time_range，如果没有则使用 interval
+            time_range_value = query.time_range if query.time_range else query.interval
+            delta = time_range_map.get(time_range_value, timedelta(minutes=15))
+            
+            # 获取队列的最新任务时间，确保图表包含最新数据
+            queue_name = query.queues[0] if query.queues else None
+            if queue_name:
+                latest_time = await data_access.get_latest_task_time(queue_name)
+                if latest_time:
+                    # 使用最新任务时间作为结束时间
+                    end_time = latest_time.replace(second=59, microsecond=999999)  # 包含整分钟
+                    logger.info(f"使用最新任务时间: {latest_time}")
+                else:
+                    # 如果没有任务，使用当前时间
+                    end_time = now.replace(second=0, microsecond=0)
+            else:
+                end_time = now.replace(second=0, microsecond=0)
+            
+            start_time = end_time - delta
+            logger.info(f"使用预设时间范围 {time_range_value}: {start_time} 到 {end_time}, delta: {delta}")
+        
+        # 确保有队列名称
+        if not query.queues or len(query.queues) == 0:
+            return {"data": [], "granularity": "minute"}
+        
+        # 获取第一个队列的流量速率
+        queue_name = query.queues[0]
+        # TimeRangeQuery 没有 filters 属性，传递 None 或空字典
+        filters = getattr(query, 'filters', None)
+        data, granularity = await data_access.fetch_queue_flow_rates(
+            queue_name, start_time, end_time, filters
+        )
+        
+        return {"data": data, "granularity": granularity}
+    
+    @staticmethod
+    async def get_global_stats(data_access) -> Dict[str, Any]:
+        """
+        获取全局统计信息
+        
+        Args:
+            data_access: 数据访问层实例
+            
+        Returns:
+            全局统计数据
+        """
+        stats_data = await data_access.fetch_global_stats()
+        return {
+            "success": True,
+            "data": stats_data
+        }
+    
+    @staticmethod
+    async def get_queues_detail(data_access) -> Dict[str, Any]:
+        """
+        获取队列详细信息
+        
+        Args:
+            data_access: 数据访问层实例
+            
+        Returns:
+            队列详细数据
+        """
+        queues_data = await data_access.fetch_queues_data()
+        return {
+            "success": True,
+            "data": queues_data
+        }
+    
+    @staticmethod
+    async def delete_queue(queue_name: str) -> Dict[str, Any]:
+        """
+        删除队列
+        
+        Args:
+            queue_name: 队列名称
+            
+        Returns:
+            操作结果
+        """
+        # TODO: 实现删除队列的逻辑
+        logger.info(f"删除队列请求: {queue_name}")
+        return {
+            "success": True,
+            "message": f"队列 {queue_name} 已删除"
+        }
+    
+    @staticmethod
+    async def trim_queue(queue_name: str, max_length: int) -> Dict[str, Any]:
+        """
+        裁剪队列到指定长度
+        
+        Args:
+            queue_name: 队列名称
+            max_length: 最大长度
+            
+        Returns:
+            操作结果
+        """
+        # TODO: 实现裁剪队列的逻辑
+        logger.info(f"裁剪队列请求: {queue_name}, 保留 {max_length} 条消息")
+        return {
+            "success": True,
+            "message": f"队列 {queue_name} 已裁剪至 {max_length} 条消息"
+        }
+    
+    @staticmethod
+    async def get_queue_stats_v2(
+        namespace_data_access,
+        namespace: str,
+        queue: Optional[str] = None,
+        start_time: Optional[datetime] = None,
+        end_time: Optional[datetime] = None,
+        time_range: Optional[str] = None
+    ) -> Dict[str, Any]:
+        """
+        获取队列统计信息v2 - 支持消费者组详情和优先级队列
+        
+        Args:
+            namespace_data_access: 命名空间数据访问实例
+            namespace: 命名空间
+            queue: 可选，筛选特定队列
+            start_time: 开始时间
+            end_time: 结束时间
+            time_range: 时间范围
+            
+        Returns:
+            队列统计数据
+        """
+        # 获取命名空间连接
+        conn = await namespace_data_access.manager.get_connection(namespace)
+        
+        # 获取Redis客户端
+        redis_client = await conn.get_redis_client(decode=False)
+        
+        # 获取PostgreSQL会话（可选）
+        pg_session = None
+        if conn.AsyncSessionLocal:
+            pg_session = conn.AsyncSessionLocal()
+        
+        try:
+            # 导入 QueueStatsV2
+            from jettask.backend.queue_stats_v2 import QueueStatsV2
+            
+            # 创建统计服务实例
+            stats_service = QueueStatsV2(
+                redis_client=redis_client,
+                pg_session=pg_session,
+                redis_prefix=conn.redis_prefix
+            )
+            
+            # 处理时间筛选参数
+            time_filter = None
+            if time_range or start_time or end_time:
+                time_filter = {}
+                
+                # 如果提供了time_range，计算开始和结束时间
+                if time_range and time_range != 'custom':
+                    now = datetime.now(timezone.utc)
+                    if time_range.endswith('m'):
+                        minutes = int(time_range[:-1])
+                        time_filter['start_time'] = now - timedelta(minutes=minutes)
+                        time_filter['end_time'] = now
+                    elif time_range.endswith('h'):
+                        hours = int(time_range[:-1])
+                        time_filter['start_time'] = now - timedelta(hours=hours)
+                        time_filter['end_time'] = now
+                    elif time_range.endswith('d'):
+                        days = int(time_range[:-1])
+                        time_filter['start_time'] = now - timedelta(days=days)
+                        time_filter['end_time'] = now
+                else:
+                    # 使用提供的start_time和end_time
+                    if start_time:
+                        time_filter['start_time'] = start_time
+                    if end_time:
+                        time_filter['end_time'] = end_time
+            
+            # 获取队列统计（使用分组格式）
+            stats = await stats_service.get_queue_stats_grouped(time_filter)
+            
+            # 如果指定了队列筛选，则过滤结果
+            if queue:
+                stats = [s for s in stats if s['queue_name'] == queue]
+            
+            return {
+                "success": True,
+                "data": stats
+            }
+            
+        finally:
+            if pg_session:
+                await pg_session.close()
+            await redis_client.aclose()
+    
+    @staticmethod
+    async def get_tasks_v2(namespace_data_access, namespace: str, body: Dict[str, Any]) -> Dict[str, Any]:
+        """
+        获取任务列表v2 - 支持tasks和task_runs表连表查询
+        
+        Args:
+            namespace_data_access: 命名空间数据访问实例
+            namespace: 命名空间
+            body: 请求体参数
+            
+        Returns:
+            任务列表数据
+        """
+        from sqlalchemy import text
+        from datetime import datetime, timezone, timedelta
+        
+        queue_name = body.get('queue_name')
+        page = body.get('page', 1)
+        page_size = body.get('page_size', 20)
+        filters = body.get('filters', [])
+        time_range = body.get('time_range', '1h')
+        start_time = body.get('start_time')
+        end_time = body.get('end_time')
+        sort_field = body.get('sort_field', 'created_at')
+        sort_order = body.get('sort_order', 'desc')
+        
+        if not queue_name:
+            raise ValueError("queue_name is required")
+        
+        # 获取命名空间连接
+        conn = await namespace_data_access.manager.get_connection(namespace)
+        
+        if not conn.pg_config or not conn.async_engine:
+            return {
+                "success": True,
+                "data": [],
+                "total": 0
+            }
+        
+        # 解析时间范围
+        if start_time and end_time:
+            # 使用自定义时间范围
+            start_dt = datetime.fromisoformat(start_time.replace('Z', '+00:00'))
+            end_dt = datetime.fromisoformat(end_time.replace('Z', '+00:00'))
+        else:
+            # 使用预定义时间范围
+            end_dt = datetime.now(timezone.utc)
+            time_deltas = {
+                '15m': timedelta(minutes=15),
+                '30m': timedelta(minutes=30),
+                '1h': timedelta(hours=1),
+                '3h': timedelta(hours=3),
+                '6h': timedelta(hours=6),
+                '12h': timedelta(hours=12),
+                '1d': timedelta(days=1),
+                '3d': timedelta(days=3),
+                '7d': timedelta(days=7),
+                '30d': timedelta(days=30)
+            }
+            delta = time_deltas.get(time_range, timedelta(hours=1))
+            start_dt = end_dt - delta
+        
+        offset = (page - 1) * page_size
+        
+        async with conn.async_engine.begin() as pg_conn:
+            # 构建查询条件
+            conditions = [
+                "t.namespace = :namespace",
+                "t.queue = :queue",
+                "t.created_at >= :start_time",
+                "t.created_at <= :end_time"
+            ]
+            query_params = {
+                "namespace": namespace,
+                "queue": queue_name,
+                "start_time": start_dt,
+                "end_time": end_dt,
+                "limit": page_size,
+                "offset": offset
+            }
+            
+            # 处理筛选条件
+            for i, filter_item in enumerate(filters):
+                field = filter_item.get('field')
+                operator = filter_item.get('operator')
+                value = filter_item.get('value')
+                
+                if field and operator and value is not None:
+                    param_key = f"filter_{i}"
+                    
+                    # 映射前端字段到数据库字段（使用payload JSONB列）
+                    db_field_map = {
+                        'id': 't.stream_id',
+                        'task_name': "t.payload::jsonb->'event_data'->>'__task_name'",
+                        'status': "t.payload::jsonb->>'status'",
+                        'worker_id': "t.payload::jsonb->>'worker_id'",
+                        'scheduled_task_id': 't.scheduled_task_id'
+                    }
+                    
+                    db_field = db_field_map.get(field, f't.{field}')
+                    
+                    if operator == 'eq':
+                        conditions.append(f"{db_field} = :{param_key}")
+                        query_params[param_key] = value
+                    elif operator == 'contains':
+                        conditions.append(f"{db_field} LIKE :{param_key}")
+                        query_params[param_key] = f"%{value}%"
+                    elif operator == 'gt':
+                        conditions.append(f"{db_field} > :{param_key}")
+                        query_params[param_key] = value
+                    elif operator == 'lt':
+                        conditions.append(f"{db_field} < :{param_key}")
+                        query_params[param_key] = value
+            
+            where_clause = " AND ".join(conditions)
+            
+            # 获取总数
+            count_query = f"""
+                SELECT COUNT(*) as total 
+                FROM tasks t
+                WHERE {where_clause}
+            """
+            count_result = await pg_conn.execute(text(count_query), query_params)
+            total = count_result.fetchone().total
+            
+            # 构建排序
+            sort_map = {
+                'created_at': 't.created_at',
+                'started_at': 't.started_at',
+                'completed_at': 't.completed_at'
+            }
+            order_by = sort_map.get(sort_field, 't.created_at')
+            order_direction = 'DESC' if sort_order == 'desc' else 'ASC'
+            
+            # 获取任务列表（从payload JSONB中提取数据）
+            query = f"""
+                SELECT 
+                    t.stream_id as id,
+                    t.payload::jsonb->'event_data'->>'__task_name' as task_name,
+                    t.queue,
+                    t.payload::jsonb->>'status' as status,
+                    t.priority,
+                    COALESCE((t.payload::jsonb->>'retry_count')::int, 0) as retry_count,
+                    COALESCE((t.payload::jsonb->>'max_retry')::int, 3) as max_retry,
+                    t.created_at,
+                    (t.payload::jsonb->>'started_at')::timestamptz as started_at,
+                    (t.payload::jsonb->>'completed_at')::timestamptz as completed_at,
+                    t.payload::jsonb->>'worker_id' as worker_id,
+                    t.payload::jsonb->>'error_message' as error_message,
+                    (t.payload::jsonb->>'execution_time')::float as execution_time,
+                    CASE 
+                        WHEN t.payload::jsonb->>'completed_at' IS NOT NULL AND t.created_at IS NOT NULL 
+                        THEN EXTRACT(EPOCH FROM ((t.payload::jsonb->>'completed_at')::timestamptz - t.created_at))
+                        ELSE NULL 
+                    END as duration,
+                    t.scheduled_task_id,
+                    t.source,
+                    t.metadata
+                FROM tasks t
+                WHERE {where_clause}
+                ORDER BY {order_by} {order_direction}
+                LIMIT :limit OFFSET :offset
+            """
+            
+            result = await pg_conn.execute(text(query), query_params)
+            
+            tasks = []
+            for row in result:
+                tasks.append({
+                    "id": row.id,
+                    "task_name": row.task_name or "unknown",
+                    "queue": row.queue,
+                    "status": row.status,
+                    "priority": row.priority,
+                    "retry_count": row.retry_count,
+                    "max_retry": row.max_retry,
+                    "created_at": row.created_at.isoformat() if row.created_at else None,
+                    "started_at": row.started_at.isoformat() if row.started_at else None,
+                    "completed_at": row.completed_at.isoformat() if row.completed_at else None,
+                    "duration": round(row.duration, 2) if row.duration else None,
+                    "execution_time": float(row.execution_time) if row.execution_time else None,
+                    "worker_id": row.worker_id,
+                    "error_message": row.error_message
+                })
+            
+            return {
+                "success": True,
+                "data": tasks,
+                "total": total
+            }
+    
+    @staticmethod
+    async def get_consumer_group_stats(namespace_data_access, namespace: str, group_name: str) -> Dict[str, Any]:
+        """
+        获取特定消费者组的详细统计
+        
+        Args:
+            namespace_data_access: 命名空间数据访问实例
+            namespace: 命名空间
+            group_name: 消费者组名称
+            
+        Returns:
+            消费者组统计数据
+        """
+        # 获取命名空间连接
+        conn = await namespace_data_access.manager.get_connection(namespace)
+        
+        # 获取PostgreSQL会话
+        if not conn.AsyncSessionLocal:
+            raise ValueError("PostgreSQL not configured for this namespace")
+        
+        async with conn.AsyncSessionLocal() as session:
+            # 查询消费者组的执行统计
+            query = text("""
+                WITH group_stats AS (
+                    SELECT 
+                        tr.consumer_group,
+                        tr.task_name,
+                        COUNT(*) as total_tasks,
+                        COUNT(CASE WHEN tr.status = 'success' THEN 1 END) as success_count,
+                        COUNT(CASE WHEN tr.status = 'failed' THEN 1 END) as failed_count,
+                        COUNT(CASE WHEN tr.status = 'running' THEN 1 END) as running_count,
+                        AVG(tr.execution_time) as avg_execution_time,
+                        MIN(tr.execution_time) as min_execution_time,
+                        MAX(tr.execution_time) as max_execution_time,
+                        AVG(tr.duration) as avg_duration,
+                        MIN(tr.started_at) as first_task_time,
+                        MAX(tr.completed_at) as last_task_time
+                    FROM task_runs tr
+                    WHERE tr.consumer_group = :group_name
+                        AND tr.started_at > NOW() - INTERVAL '24 hours'
+                    GROUP BY tr.consumer_group, tr.task_name
+                ),
+                hourly_stats AS (
+                    SELECT 
+                        DATE_TRUNC('hour', tr.started_at) as hour,
+                        COUNT(*) as task_count,
+                        AVG(tr.execution_time) as avg_exec_time
+                    FROM task_runs tr
+                    WHERE tr.consumer_group = :group_name
+                        AND tr.started_at > NOW() - INTERVAL '24 hours'
+                    GROUP BY DATE_TRUNC('hour', tr.started_at)
+                    ORDER BY hour
+                )
+                SELECT 
+                    (SELECT row_to_json(gs) FROM group_stats gs) as summary,
+                    (SELECT json_agg(hs) FROM hourly_stats hs) as hourly_trend
+            """)
+            
+            result = await session.execute(query, {'group_name': group_name})
+            row = result.fetchone()
+            
+            if not row or not row.summary:
+                return {
+                    "success": True,
+                    "data": {
+                        "group_name": group_name,
+                        "summary": {},
+                        "hourly_trend": []
+                    }
+                }
+            
+            return {
+                "success": True,
+                "data": {
+                    "group_name": group_name,
+                    "summary": row.summary,
+                    "hourly_trend": row.hourly_trend or []
+                }
+            }
+    
+    @staticmethod
+    async def get_stream_backlog(
+        data_access,
+        namespace: str,
+        stream_name: Optional[str] = None,
+        hours: int = 24
+    ) -> Dict[str, Any]:
+        """
+        获取Stream积压监控数据
+        
+        Args:
+            data_access: 数据访问层实例
+            namespace: 命名空间
+            stream_name: 可选，指定stream名称
+            hours: 查询最近多少小时的数据
+            
+        Returns:
+            Stream积压数据
+        """
+        # 计算时间范围
+        end_time = datetime.now(timezone.utc)
+        start_time = end_time - timedelta(hours=hours)
+        
+        async with data_access.AsyncSessionLocal() as session:
+            # 构建查询
+            if stream_name:
+                query = text("""
+                    SELECT 
+                        stream_name,
+                        consumer_group,
+                        last_published_offset,
+                        last_delivered_offset,
+                        last_acked_offset,
+                        pending_count,
+                        backlog_undelivered,
+                        backlog_unprocessed,
+                        created_at
+                    FROM stream_backlog_monitor
+                    WHERE namespace = :namespace
+                        AND stream_name = :stream_name
+                        AND created_at >= :start_time
+                        AND created_at <= :end_time
+                    ORDER BY created_at DESC
+                    LIMIT 1000
+                """)
+                params = {
+                    'namespace': namespace,
+                    'stream_name': stream_name,
+                    'start_time': start_time,
+                    'end_time': end_time
+                }
+            else:
+                # 获取最新的所有stream数据
+                query = text("""
+                    SELECT DISTINCT ON (stream_name, consumer_group)
+                        stream_name,
+                        consumer_group,
+                        last_published_offset,
+                        last_delivered_offset,
+                        last_acked_offset,
+                        pending_count,
+                        backlog_undelivered,
+                        backlog_unprocessed,
+                        created_at
+                    FROM stream_backlog_monitor
+                    WHERE namespace = :namespace
+                        AND created_at >= :start_time
+                    ORDER BY stream_name, consumer_group, created_at DESC
+                """)
+                params = {
+                    'namespace': namespace,
+                    'start_time': start_time
+                }
+            
+            result = await session.execute(query, params)
+            rows = result.fetchall()
+            
+            # 格式化数据
+            data = []
+            for row in rows:
+                data.append({
+                    'stream_name': row.stream_name,
+                    'consumer_group': row.consumer_group,
+                    'last_published_offset': row.last_published_offset,
+                    'last_delivered_offset': row.last_delivered_offset,
+                    'last_acked_offset': row.last_acked_offset,
+                    'pending_count': row.pending_count,
+                    'backlog_undelivered': row.backlog_undelivered,
+                    'backlog_unprocessed': row.backlog_unprocessed,
+                    'created_at': row.created_at.isoformat() if row.created_at else None
+                })
+            
+            return {
+                'success': True,
+                'data': data,
+                'total': len(data)
+            }
+    
+    @staticmethod
+    async def get_stream_backlog_summary(data_access, namespace: str) -> Dict[str, Any]:
+        """
+        获取Stream积压监控汇总数据
+        
+        Args:
+            data_access: 数据访问层实例
+            namespace: 命名空间
+            
+        Returns:
+            汇总数据
+        """
+        async with data_access.AsyncSessionLocal() as session:
+            # 获取最新的汇总数据
+            query = text("""
+                WITH latest_data AS (
+                    SELECT DISTINCT ON (stream_name, consumer_group)
+                        stream_name,
+                        consumer_group,
+                        backlog_undelivered,
+                        backlog_unprocessed,
+                        pending_count
+                    FROM stream_backlog_monitor
+                    WHERE namespace = :namespace
+                        AND created_at >= NOW() - INTERVAL '1 hour'
+                    ORDER BY stream_name, consumer_group, created_at DESC
+                )
+                SELECT 
+                    COUNT(DISTINCT stream_name) as total_streams,
+                    COUNT(DISTINCT consumer_group) as total_groups,
+                    SUM(backlog_unprocessed) as total_backlog,
+                    SUM(pending_count) as total_pending,
+                    MAX(backlog_unprocessed) as max_backlog
+                FROM latest_data
+            """)
+            
+            result = await session.execute(query, {'namespace': namespace})
+            row = result.fetchone()
+            
+            if row:
+                return {
+                    'success': True,
+                    'data': {
+                        'total_streams': row.total_streams or 0,
+                        'total_groups': row.total_groups or 0,
+                        'total_backlog': row.total_backlog or 0,
+                        'total_pending': row.total_pending or 0,
+                        'max_backlog': row.max_backlog or 0
+                    }
+                }
+            else:
+                return {
+                    'success': True,
+                    'data': {
+                        'total_streams': 0,
+                        'total_groups': 0,
+                        'total_backlog': 0,
+                        'total_pending': 0,
+                        'max_backlog': 0
+                    }
+                }