Fix: check external commands after node completion (#26891)

Co-authored-by: crazywoola <100913391+crazywoola@users.noreply.github.com>

api/core/workflow/graph_engine/command_channels/redis_channel.py

@@ -41,6 +41,7 @@ class RedisChannel:
         self._redis = redis_client
         self._key = channel_key
         self._command_ttl = command_ttl
+        self._pending_key = f"{channel_key}:pending"
 
     def fetch_commands(self) -> list[GraphEngineCommand]:
         """
@@ -49,6 +50,9 @@ class RedisChannel:
         Returns:
             List of pending commands (drains the Redis list)
         """
+        if not self._has_pending_commands():
+            return []
+
         commands: list[GraphEngineCommand] = []
 
         # Use pipeline for atomic operations
@@ -85,6 +89,7 @@ class RedisChannel:
         with self._redis.pipeline() as pipe:
             pipe.rpush(self._key, command_json)
             pipe.expire(self._key, self._command_ttl)
+            pipe.set(self._pending_key, "1", ex=self._command_ttl)
             pipe.execute()
 
     def _deserialize_command(self, data: dict[str, Any]) -> GraphEngineCommand | None:
@@ -112,3 +117,17 @@ class RedisChannel:
 
         except (ValueError, TypeError):
             return None
+
+    def _has_pending_commands(self) -> bool:
+        """
+        Check and consume the pending marker to avoid unnecessary list reads.
+
+        Returns:
+            True if commands should be fetched from Redis.
+        """
+        with self._redis.pipeline() as pipe:
+            pipe.get(self._pending_key)
+            pipe.delete(self._pending_key)
+            pending_value, _ = pipe.execute()
+
+        return pending_value is not None

api/core/workflow/graph_engine/orchestration/dispatcher.py

@@ -8,7 +8,12 @@ import threading
 import time
 from typing import TYPE_CHECKING, final
 
-from core.workflow.graph_events.base import GraphNodeEventBase
+from core.workflow.graph_events import (
+    GraphNodeEventBase,
+    NodeRunExceptionEvent,
+    NodeRunFailedEvent,
+    NodeRunSucceededEvent,
+)
 
 from ..event_management import EventManager
 from .execution_coordinator import ExecutionCoordinator
@@ -72,13 +77,16 @@ class Dispatcher:
         if self._thread and self._thread.is_alive():
             self._thread.join(timeout=10.0)
 
+    _COMMAND_TRIGGER_EVENTS = (
+        NodeRunSucceededEvent,
+        NodeRunFailedEvent,
+        NodeRunExceptionEvent,
+    )
+
     def _dispatcher_loop(self) -> None:
         """Main dispatcher loop."""
         try:
             while not self._stop_event.is_set():
-                # Check for commands
-                self._execution_coordinator.check_commands()
-
                 # Check for scaling
                 self._execution_coordinator.check_scaling()
 
@@ -87,6 +95,8 @@ class Dispatcher:
                     event = self._event_queue.get(timeout=0.1)
                     # Route to the event handler
                     self._event_handler.dispatch(event)
+                    if self._should_check_commands(event):
+                        self._execution_coordinator.check_commands()
                     self._event_queue.task_done()
                 except queue.Empty:
                     # Check if execution is complete
@@ -102,3 +112,7 @@ class Dispatcher:
             # Signal the event emitter that execution is complete
             if self._event_emitter:
                 self._event_emitter.mark_complete()
+
+    def _should_check_commands(self, event: GraphNodeEventBase) -> bool:
+        """Return True if the event represents a node completion."""
+        return isinstance(event, self._COMMAND_TRIGGER_EVENTS)