UN-2813 [FIX] Implement graceful shutdown for signal handler

Improve signal handler to properly shut down backend instead of
immediate exit that can leave tasks in inconsistent state.

**Problem:**
Signal handler immediately called sys.exit(0) without giving backend
a chance to finish current tasks or clean up resources:
```python
def _signal_handler(self, signum, frame):
    logger.info("Initiating worker shutdown")
    sys.exit(0)  # ❌ Immediate exit, no cleanup
```

This can cause:
- Tasks left in inconsistent states
- Database connections not closed
- Incomplete work lost
- No proper cleanup of backend resources

**Solution:**
Implement graceful shutdown that:
1. Sets shutdown_requested flag
2. Calls backend.stop(graceful=True) if available
3. Logs errors with full traceback
4. Allows backend to finish current tasks

```python
def _signal_handler(self, signum, frame):
    self.shutdown_requested = True
    logger.info("Initiating graceful worker shutdown")
    if self.backend and hasattr(self.backend, "stop"):
        try:
            self.backend.stop(graceful=True)
        except Exception as e:
            logger.error(f"Error during graceful shutdown: {e}", exc_info=True)
    sys.exit(0)
```

**Changes:**
1. ✅ Added backend.stop(graceful=True) call before exit
2. ✅ Proper try/except with exc_info logging
3. ✅ Check for backend availability with hasattr
4. ✅ More descriptive log messages
5. ✅ shutdown_requested flag still set for future use

**Impact:**
- ✅ Tasks can complete gracefully before shutdown
- ✅ Backend resources properly cleaned up
- ✅ Better error visibility during shutdown
- ✅ Safer production deployments

**Note:**
Backend implementations should support stop() method for proper
graceful shutdown:
- Stop accepting new tasks
- Wait for in-progress tasks (with timeout)
- Clean up connections and resources
- Exit cleanly

Resolves: CodeRabbit comment #2404866807
Related: #1555

Author: Deepak-Kesavan

task-backend/src/unstract/task_backend/worker.py

@@ -223,13 +223,19 @@ def _get_worker_config(self) -> dict:
         return base_config
 
     def _signal_handler(self, signum, frame) -> None:
-        """Handle shutdown signals."""
+        """Handle shutdown signals gracefully."""
         logger.info(f"Received shutdown signal: {signum}")
         self.shutdown_requested = True
 
-        # Note: The backend's run_worker() method should handle
-        # graceful shutdown. For now, we'll just exit.
-        logger.info("Initiating worker shutdown")
+        logger.info("Initiating graceful worker shutdown")
+        if self.backend and hasattr(self.backend, "stop"):
+            try:
+                # Give the backend a chance to finish current tasks
+                logger.info("Requesting backend graceful shutdown")
+                self.backend.stop(graceful=True)
+            except Exception as e:
+                logger.error(f"Error during graceful shutdown: {e}", exc_info=True)
+
         sys.exit(0)