Index rebuild taking much time

Can you provide the full output of the cluster stats API? How much heap do you data nodes have assigned?