Reduce 1 worker

app.py

@@ -17,12 +17,10 @@ The chatbot instructs Gemini Flash to format its answer using markdown.
 import os
 import faiss
 import numpy as np
-import gc
 import time
 from fastapi import FastAPI
 from fastapi.responses import JSONResponse
 from pathlib import Path
-# import requests
 import threading
 from dotenv import load_dotenv
 
@@ -122,9 +120,9 @@ else:
     print("✅ Loaded existing QA data from MongoDB.")
     # Use an aggregation pipeline with allowDiskUse to sort by "i" without creating an index.
     qa_docs = list(qa_collection.aggregate([
-        {"$sort": {"i": 1}},
-        {"$project": {"_id": 0}}
-    ], allowDiskUse=True))
+            {"$sort": {"i": 1}},
+            {"$project": {"_id": 0}}
+        ], allowDiskUse=True))
     qa_data = qa_docs
     print("📦 Total QA entries loaded:", len(qa_data))
 
@@ -265,7 +263,7 @@ import uvicorn
 if __name__ == "__main__":
     try:
         print("✅ Server is starting...")
-        uvicorn.run(app, host="0.0.0.0", port=7860)
+        uvicorn.run(app, host="0.0.0.0", port=7860, workers=1) # Default 2 workers, cut to 1
     except Exception as e:
         print(f"❌ Server startup failed: {e}")
         exit(1)