I have tried reading the documents in different format, I have an array of keys to use for reading. Each batch contains 1000 keys. I have 6 riak nodes. I do reads with r=1. I connect each time to the same riak node. Documents are just profile fields, so it's not something big. I've checked the CPU and disk usage on the nodes, and I can observe just a slight move, but not even close to overall 20% CPU.I have tried reading the documents in different