refactor: streamline trigger event node metadata handling and update async workflow service for JSON serialization

- Removed unnecessary input data from the TriggerEventNode's metadata.
- Updated AsyncWorkflowService to use model_dump_json() for trigger metadata serialization.
- Added a comment in WorkflowAppService to address the large size of the workflow_app_log table and the use of an additional details field.

api/services/workflow/entities.py

@@ -19,6 +19,12 @@ class AsyncTriggerStatus(StrEnum):
     TIMEOUT = "timeout"
 
 
+class TriggerMetadata(BaseModel):
+    """Trigger metadata"""
+
+    pass
+
+
 class TriggerData(BaseModel):
     """Base trigger data model for async workflow execution"""
 
@@ -30,6 +36,7 @@ class TriggerData(BaseModel):
     files: Sequence[Mapping[str, Any]] = Field(default_factory=list)
     trigger_type: AppTriggerType
     trigger_from: WorkflowRunTriggeredFrom
+    trigger_metadata: TriggerMetadata = Field(default_factory=TriggerMetadata)
 
     model_config = ConfigDict(use_enum_values=True)
 
@@ -48,6 +55,17 @@ class ScheduleTriggerData(TriggerData):
     trigger_from: WorkflowRunTriggeredFrom = WorkflowRunTriggeredFrom.SCHEDULE
 
 
+class PluginTriggerMetadata(TriggerMetadata):
+    """Plugin trigger metadata"""
+
+    plugin_id: str
+    endpoint_id: str
+    plugin_unique_identifier: str
+    provider_id: str
+    icon_url: str
+    icon_dark_url: str
+
+
 class PluginTriggerData(TriggerData):
     """Plugin webhook trigger data"""