Handling 500—599 server errors

[rusq]

Describe the bug
Currently we crash and burn on 500, and this is an issue on a long running jobs, as there's no retry functionality (to be fixed in scope of #172).

To Reproduce
Steps to reproduce the behavior:

1. Run Slackdump Export on a large workspace
2. Depending on your luck you might get 500 Server Error and Slackdump will terminate.

Expected behavior
A clear and concise retry on 500-599 errors.

References:

1. Resume-from-failed functionality #174