diff --git a/jina/executors/indexers/__init__.py b/jina/executors/indexers/__init__.py
index f54aa3c7589e9..17a6d0f29e208 100644
--- a/jina/executors/indexers/__init__.py
+++ b/jina/executors/indexers/__init__.py
@@ -233,4 +233,3 @@ class ChunkIndexer(CompoundExecutor):
             - !ControlReqDriver {}
     """
 
-    pass
diff --git a/jina/executors/indexers/keyvalue/proto.py b/jina/executors/indexers/keyvalue/proto.py
index aa0b0dc111b14..b240d076b3c62 100644
--- a/jina/executors/indexers/keyvalue/proto.py
+++ b/jina/executors/indexers/keyvalue/proto.py
@@ -15,6 +15,7 @@ class BasePbIndexer(BaseKVIndexer):
     """Storing and querying protobuf chunk/document using gzip and Python dict. """
 
     compress_level = 1  #: The compresslevel argument is an integer from 0 to 9 controlling the level of compression
+    flush_on_add = True  #: When set to true, the indexer is flushed on every add, it is safer but slower
 
     def get_query_handler(self):
         r = {}
@@ -47,7 +48,8 @@ def add(self, obj):
         """
         json.dump(obj, self.write_handler)
         self.write_handler.write('\n')
-        self.flush()
+        if self.flush_on_add:
+            self.flush()
 
     def query(self, key: int) -> Union['jina_pb2.Chunk', 'jina_pb2.Document']:
         """ Find the protobuf chunk/doc using id