try(hotpatch): add autoscaling command

surfsense_backend/app/config/global_llm_config.example.yaml

@@ -122,8 +122,52 @@ global_llm_configs:
     use_default_system_instructions: false
     citations_enabled: true
 
-  # Example: Groq - Fast inference
+  # Example: Azure OpenAI GPT-4o
+  # IMPORTANT: For Azure deployments, always include 'base_model' in litellm_params
+  # to enable accurate token counting, cost tracking, and max token limits
   - id: -5
+    name: "Global Azure GPT-4o"
+    description: "Azure OpenAI GPT-4o deployment"
+    provider: "AZURE"
+    # model_name format for Azure: azure/<your-deployment-name>
+    model_name: "azure/gpt-4o-deployment"
+    api_key: "your-azure-api-key-here"
+    api_base: "https://your-resource.openai.azure.com"
+    api_version: "2024-02-15-preview"  # Azure API version
+    rpm: 1000
+    tpm: 150000
+    litellm_params:
+      temperature: 0.7
+      max_tokens: 4000
+      # REQUIRED for Azure: Specify the underlying OpenAI model
+      # This fixes "Could not identify azure model" warnings
+      # Common base_model values: gpt-4, gpt-4-turbo, gpt-4o, gpt-4o-mini, gpt-3.5-turbo
+      base_model: "gpt-4o"
+    system_instructions: ""
+    use_default_system_instructions: true
+    citations_enabled: true
+
+  # Example: Azure OpenAI GPT-4 Turbo
+  - id: -6
+    name: "Global Azure GPT-4 Turbo"
+    description: "Azure OpenAI GPT-4 Turbo deployment"
+    provider: "AZURE"
+    model_name: "azure/gpt-4-turbo-deployment"
+    api_key: "your-azure-api-key-here"
+    api_base: "https://your-resource.openai.azure.com"
+    api_version: "2024-02-15-preview"
+    rpm: 500
+    tpm: 100000
+    litellm_params:
+      temperature: 0.7
+      max_tokens: 4000
+      base_model: "gpt-4-turbo"  # Maps to gpt-4-turbo-preview
+    system_instructions: ""
+    use_default_system_instructions: true
+    citations_enabled: true
+
+  # Example: Groq - Fast inference
+  - id: -7
     name: "Global Groq Llama 3"
     description: "Ultra-fast Llama 3 70B via Groq"
     provider: "GROQ"
@@ -150,3 +194,11 @@ global_llm_configs:
 # - All standard LiteLLM providers are supported
 # - rpm/tpm: Optional rate limits for load balancing (requests/tokens per minute)
 #   These help the router distribute load evenly and avoid rate limit errors
+#
+# AZURE-SPECIFIC NOTES:
+# - Always add 'base_model' in litellm_params for Azure deployments
+# - This fixes "Could not identify azure model 'X'" warnings
+# - base_model should match the underlying OpenAI model (e.g., gpt-4o, gpt-4-turbo, gpt-3.5-turbo)
+# - model_name format: "azure/<your-deployment-name>"
+# - api_version: Use a recent Azure API version (e.g., "2024-02-15-preview")
+# - See: https://docs.litellm.ai/docs/proxy/cost_tracking#spend-tracking-for-azure-openai-models

surfsense_backend/app/tasks/celery_tasks/document_tasks.py

@@ -323,6 +323,28 @@ def process_file_upload_task(
         user_id: ID of the user
     """
     import asyncio
+    import os
+    import traceback
+
+    logger.info(
+        f"[process_file_upload] Task started - file: {filename}, "
+        f"search_space_id: {search_space_id}, user_id: {user_id}"
+    )
+    logger.info(f"[process_file_upload] File path: {file_path}")
+
+    # Check if file exists and is accessible
+    if not os.path.exists(file_path):
+        logger.error(
+            f"[process_file_upload] File does not exist: {file_path}. "
+            "The temp file may have been cleaned up before the task ran."
+        )
+        return
+
+    try:
+        file_size = os.path.getsize(file_path)
+        logger.info(f"[process_file_upload] File size: {file_size} bytes")
+    except Exception as e:
+        logger.warning(f"[process_file_upload] Could not get file size: {e}")
 
     loop = asyncio.new_event_loop()
     asyncio.set_event_loop(loop)
@@ -331,6 +353,13 @@ def process_file_upload_task(
         loop.run_until_complete(
             _process_file_upload(file_path, filename, search_space_id, user_id)
         )
+        logger.info(f"[process_file_upload] Task completed successfully for: {filename}")
+    except Exception as e:
+        logger.error(
+            f"[process_file_upload] Task failed for {filename}: {e}\n"
+            f"Traceback:\n{traceback.format_exc()}"
+        )
+        raise
     finally:
         loop.close()
 
@@ -343,16 +372,22 @@ async def _process_file_upload(
 
     from app.tasks.document_processors.file_processors import process_file_in_background
 
+    logger.info(f"[_process_file_upload] Starting async processing for: {filename}")
+
     async with get_celery_session_maker()() as session:
+        logger.info(f"[_process_file_upload] Database session created for: {filename}")
         task_logger = TaskLoggingService(session, search_space_id)
 
         # Get file size for notification metadata
         try:
             file_size = os.path.getsize(file_path)
-        except Exception:
+            logger.info(f"[_process_file_upload] File size: {file_size} bytes")
+        except Exception as e:
+            logger.warning(f"[_process_file_upload] Could not get file size: {e}")
             file_size = None
 
         # Create notification for document processing
+        logger.info(f"[_process_file_upload] Creating notification for: {filename}")
         notification = (
             await NotificationService.document_processing.notify_processing_started(
                 session=session,
@@ -363,6 +398,9 @@ async def _process_file_upload(
                 file_size=file_size,
             )
         )
+        logger.info(
+            f"[_process_file_upload] Notification created with ID: {notification.id if notification else 'None'}"
+        )
 
         log_entry = await task_logger.log_task_start(
             task_name="process_file_upload",

surfsense_backend/scripts/docker/entrypoint.sh

@@ -39,7 +39,7 @@ backend_pid=$!
 sleep 5
 
 echo "Starting Celery Worker..."
-celery -A app.celery_app worker --loglevel=info &
+celery -A app.celery_app worker --loglevel=info --autoscale=64,4 &
 celery_worker_pid=$!
 
 # Wait a bit for worker to initialize