Checking onboardin flow - fix 3

services/training/app/api/training.py

@@ -83,13 +83,11 @@ async def start_training_job(
         except Exception as e:
             logger.warning("Failed to publish job started event", error=str(e))
 
-        # Start training in background
         background_tasks.add_task(
-            training_service.execute_training_job,
-            db, # Pass the database session
-            job.job_id,
-            job.tenant_id,
-            request # Pass the request object
+            training_service.execute_training_job_simple,
+            new_job_id,
+            tenant_id_str,
+            request
         )
 
         logger.info("Training job created",

services/training/app/services/training_service.py

@@ -31,6 +31,36 @@ class TrainingService:
     
     def __init__(self):
         self.ml_trainer = BakeryMLTrainer()
+        
+    async def execute_training_job_simple(self, job_id: str, tenant_id_str: str, request: TrainingJobRequest):
+        """Simple wrapper that creates its own database session"""
+        try:
+            # Import database_manager locally to avoid circular imports
+            from app.core.database import database_manager
+            
+            logger.info(f"Starting background training job {job_id} for tenant {tenant_id_str}")
+            
+            # Create new session for background task
+            async with database_manager.async_session_local() as session:
+                await self.execute_training_job(session, job_id, tenant_id_str, request)
+                await session.commit()
+                
+        except Exception as e:
+            logger.error(f"Background training job {job_id} failed: {str(e)}")
+            
+            # Try to update job status to failed
+            try:
+                from app.core.database import database_manager
+                async with database_manager.async_session_local() as error_session:
+                    await self._update_job_status(
+                        error_session, job_id, "failed", 0, 
+                        f"Training failed: {str(e)}", error_message=str(e)
+                    )
+                    await error_session.commit()
+            except Exception as update_error:
+                logger.error(f"Failed to update job status: {str(update_error)}")
+            
+            raise
     
     async def create_training_job(self,
                                  db: AsyncSession,
@@ -425,7 +455,7 @@ class TrainingService:
                     params["limit"] = limit
                 
                 response = await client.get(
-                    f"{settings.DATA_SERVICE_URL}/api/v1/tenants/{tenant_id}/sales/",
+                    f"{settings.DATA_SERVICE_URL}/api/v1/tenants/{tenant_id}/sales",
                     params=params,
                     headers=headers,
                     timeout=30.0
@@ -516,7 +546,7 @@ class TrainingService:
                     params["end_date"] = request.end_date.isoformat()
                 
                 response = await client.get(
-                    f"{settings.DATA_SERVICE_URL}/tenants/{tenant_id}/traffic/historical",
+                    f"http://gateway:8000/tenants/{tenant_id}/traffic/historical",
                     params=params,
                     timeout=30.0
                 )