fix: improve maintenance and synchronize live conversations

backend/app/knowledge/crawl4ai_fallback.py

@@ -262,7 +262,6 @@ class Crawl4AIFallback:
                 # Always use a separate thread for running Crawl4AI to avoid conflicts
                 logger.debug("Event loop running, creating new loop in separate thread")
                 
-                import concurrent.futures
                 import threading
                 
                 result = [None, None, None]