Uploads should be made more robust/debuggable when it comes to handling failures.

[ttys0dev]

Some ideas that may improve upload reliability:

• We should automatically retry uploads on failures
• We should have mechanisms to mitigate thundering herd issues on retries such as having response codes that permanently cancel the retry for an upload and also ones that delay retries
• We should have an easy to access extension local upload history log that users can use to identify issues with uploads(and maybe have it link to the issue tracker so that reports go to the right place)
• The extension should push upload failure log records to a remote logging service so that issues can be more easily identified/tracked