Downloads from the Cask site often get corrupted before download completes
I have noticed that download from the Cask downloads page often gets corrupted before completing. I have noticed this multiple times, on different networks as well - also reported by customers.
Given we host our packages on S3, is there any info we can gather to find out the reason for this? What can we do to improve?
What does this mean? S3 stores objects with checksums and I've never seen it serve bad data. CloudFront's caching looks at these checksums and copies the data from S3 if they do not match. Most likely, the issue is a timeout on the client side, versus an issue with S3 or CloudFront. The download is likely truncated, versus being "corrupt" as that implies bad data was transferred.
Do you have any steps to reliably reproduce this?
There's nothing that we can do to troubleshoot other people's networks.