I have a single node cluster running on a i3.2xlarge on AWS (ES 6.8.3). I want to perform a restore from S3 (in the same region) and I am surprised that the network throughput moves between 200MB/s and 300MB/s. I tried setting
max_restore_bytes_per_sec to a very high value, but something else is limiting the throughput. I have ~500GB to restore and it takes 1h to do the full restore. Is there any other settings I can play with? Or can't I expect more because there are tons of separate files to pull?