If you read the thread, its explained pretty clearly. Seems like a random network reset 5xx which will happen rarely in any distributed system. Doesn't really have anything to do with the GitHub repo issue is in.
It's happening more than 1% of the time. That's not an satisfactory answer when AWS is measuring reliability in terms of 99.9%+.