There has been great success to use https://docs.aws.amazon.com/sdk-for-go/api/service/s3/s3manager/#NewDownloader to download objects from S3 in concurrent chunks. We should explore doing the same at least for the places where the download sizes are known to be large like with the debug symbol tarballs.

https://github.com/mongodb/mongo/blob/0a68308f0d39a928ed551f285ba72ca560c38576/evergreen/spawnhost/download_archive_dist_test_debug.py#L74