I have a MongoDB server cluster running on AWS. The set up is as follows:

1 x mongos router

3 x mongod config servers

1 x shard

     - 3x mongod replicaset

I am trying to do some maths with info stored in every document of one collection. The calculation can be long so to prevent cursor timeout I used no_cursor_timeout option in find() function. However, even with setting the batch size to 2, I am still getting cursor timeout error.