任务队列执行超时自动更为状态为失败，默认超时60秒

apps/jingrow/jingrow/services/queue.py

@@ -1,8 +1,11 @@
 import os
 import json
 import logging
-from typing import Any, Dict, List
+import time
+import asyncio
 import uuid
+from datetime import datetime
+from typing import Any, Dict, List
 
 import dramatiq
 from dramatiq.brokers.redis import RedisBroker
@@ -92,9 +95,8 @@ async def _execute_node_job(payload: Dict[str, Any]) -> Dict[str, Any]:
     job_id = payload.get("job_id")
     started_iso = None
     try:
-        import time
         started = time.time()
-        started_iso = __import__('datetime').datetime.now().isoformat()
+        started_iso = datetime.now().isoformat()
         modifier = get_logged_user(session_cookie) or 'system'
         await _push_status_to_jingrow(job_id, { 
             'status': 'started', 
@@ -103,7 +105,7 @@ async def _execute_node_job(payload: Dict[str, Any]) -> Dict[str, Any]:
         })
         result = await executor.execute_node(node_type, flow_id, context, inputs, config, session_cookie)
         ended = time.time()
-        ended_iso = __import__('datetime').datetime.now().isoformat()
+        ended_iso = datetime.now().isoformat()
         await _push_status_to_jingrow(job_id, {
             'status': 'finished' if result.get('success') else 'failed',
             'exc_info': None if result.get('success') else result.get('error'),
@@ -114,7 +116,7 @@ async def _execute_node_job(payload: Dict[str, Any]) -> Dict[str, Any]:
         })
         return result
     except Exception as e:
-        ended_iso = __import__('datetime').datetime.now().isoformat()
+        ended_iso = datetime.now().isoformat()
         modifier = get_logged_user(session_cookie) or 'system'
         await _push_status_to_jingrow(job_id, { 
             'status': 'failed', 
@@ -300,10 +302,11 @@ async def _execute_agent_job(payload: Dict[str, Any]) -> Dict[str, Any]:
         return {"success": False, "error": f"Failed to get agent detail: {str(e)}"}
 
 
-@dramatiq.actor(max_retries=3, time_limit=60_000)
+@dramatiq.actor(max_retries=0, time_limit=60_000)
 def execute_local_scheduled_job(job_json: str) -> None:
     """Worker 执行入口：job_json 为 JSON 字符串，包含 target_type 及其参数。
     target_type: agent
+    超时不重试（max_retries=0）
     """
     try:        
         init_hooks(clear_cache=False)
@@ -313,38 +316,79 @@ def execute_local_scheduled_job(job_json: str) -> None:
     try:
         payload = json.loads(job_json)
     except Exception as e:
+        logger.error(f"解析任务 JSON 失败: {e}")
         return
 
-    import asyncio
-
     if not payload.get("job_id"):
         payload["job_id"] = str(uuid.uuid4())
 
     target_type = payload.get("target_type")
+    job_id = payload["job_id"]
+    sc = payload.get("session_cookie") or (payload.get("context") or {}).get("session_cookie")
+    modifier = get_logged_user(sc) or 'system'
+    started = None
+    started_iso = None
 
-    if target_type == "agent":
+    try:
-        import time
+        if target_type == "agent":
-        started = time.time()
+            started = time.time()
-        started_iso = __import__('datetime').datetime.now().isoformat()
+            started_iso = datetime.now().isoformat()
-        sc = payload.get("session_cookie") or (payload.get("context") or {}).get("session_cookie")
+            asyncio.run(_push_status_to_jingrow(job_id, { 
-        modifier = get_logged_user(sc) or 'system'
+                "status": "started", 
-        asyncio.run(_push_status_to_jingrow(payload["job_id"], { 
+                "started_at": started_iso, 
-            "status": "started", 
+                "session_cookie": sc,
-            "started_at": started_iso, 
+                "modified_by": modifier
-            "session_cookie": sc,
+            }))
-            "modified_by": modifier
+            result = asyncio.run(_execute_agent_job(payload))
-        }))
+            ended = time.time()
-        result = asyncio.run(_execute_agent_job(payload))
+            ended_iso = datetime.now().isoformat()
-        ended = time.time()
+            asyncio.run(_push_status_to_jingrow(job_id, {
-        ended_iso = __import__('datetime').datetime.now().isoformat()
+                "status": "finished" if result.get("success") else "failed",
-        asyncio.run(_push_status_to_jingrow(payload["job_id"], {
+                "ended_at": ended_iso,
-            "status": "finished" if result.get("success") else "failed",
+                "time_taken": round(max(0.0, ended - started), 3),
-            "ended_at": ended_iso,
+                "session_cookie": sc,
-            "time_taken": round(max(0.0, ended - started), 3),
+                "modified_by": modifier,
-            "session_cookie": sc,
+            }))
-            "modified_by": modifier,
+        else:
-        }))
+            logger.warning(f"Unsupported target_type: {target_type}")
-    else:
+    except BaseException as e:
-        logger.warning(f"Unsupported target_type: {target_type}")
+        # 捕获所有异常（包括 TimeLimit 等系统异常），确保任务状态正确更新
+        ended_iso = datetime.now().isoformat()
+        time_taken = round(max(0.0, (time.time() - started) if started else 0.0), 3)
+        error_msg = str(e)
+        # 判断是否为超时异常
+        is_timeout = "timeout" in error_msg.lower() or "time limit" in error_msg.lower()
+        if is_timeout:
+            error_msg = f"任务执行超时: {error_msg}"
+        # 状态更新失败不应影响异常传播，使用新的事件循环避免与已中断的循环冲突
+        status_updated = False
+        try:
+            loop = asyncio.new_event_loop()
+            asyncio.set_event_loop(loop)
+            try:
+                loop.run_until_complete(_push_status_to_jingrow(job_id, {
+                    "status": "failed",
+                    "ended_at": ended_iso,
+                    "time_taken": time_taken,
+                    "exc_info": error_msg,
+                    "session_cookie": sc,
+                    "modified_by": modifier,
+                }))
+                status_updated = True
+            finally:
+                loop.close()
+        except Exception as update_error:
+            logger.error(f"更新任务 {job_id} 状态失败: {update_error}")
+        
+        # 状态更新成功则只记录警告，不抛出异常；更新失败才抛出异常
+        if status_updated:
+            if is_timeout:
+                logger.warning(f"任务 {job_id} 执行超时，已更新状态为失败")
+            else:
+                logger.warning(f"任务 {job_id} 执行失败: {error_msg}")
+            return
+        else:
+            logger.error(f"任务 {job_id} 执行失败且状态更新失败: {e}")
+            raise