fix: add missing await in vllm-v1 `clear_kv_blocks` endpoint (#1642)

Signed-off-by: Will Killian <wkillian@nvidia.com>

launch/dynamo-run/src/subprocess/vllm_v1_inc.py

@@ -126,7 +126,7 @@ class RequestHandler:
 
     async def clear_kv_blocks(self, request=None):
         try:
-            self.engine_client.reset_prefix_cache()
+            await self.engine_client.reset_prefix_cache()
             yield {"status": "success", "message": "KV cache cleared"}
         except Exception as e:
             yield {"status": "error", "message": str(e)}