"Bad internet" causes learnersync to terminate rather than retry

[matthewbloch]

A learner reported this issue during a session, and ended up scripting their own retry loop! Let's assume that any TCP or DNS problems are probably "bad internet" and should cause learnersync to retry instead of quit. This might not apply as easily during the initial checks where we need a positive response to check the attendance_id. But after that, bad internet should be a possibility.