We are hosting one of our binary packages on Bintray in a private repository and give user a signed URL when downloading from our website.
If we open Bintray download statistics (live log) we see really strange records for one and the same file (it is our normal file):
time IP file size user
1500912829000 114.4.79.235 /bla-bla.exe 72016 anonymous
1500912828000 114.4.79.235 /bla-bla.exe 56756 anonymous
1500912828000 114.4.79.235 /bla-bla.exe 24049 anonymous
...
A lot of downloads with the same IP and different file size.
It seems that Bintray counts partial downloads as a unique download attempt. When we open statistics graphs we see really big numbers of downloads, but now we assume that these numbers are fake.
Does anybody know how Bintray counts partial downloads?
Bintray displays partial download transactions in downloads statistics, since there is no way of reliably telling if multiple partial downloads from a single origin amount to a full download/s.
The total bytes consumed by partial downloads against your account in calculated correctly, however.
One possible explanation to what you are seeing is customers using a download-manager browser extension.
Disclaimer, I work for JFrog, the company behind Bintray.