I'd like to have a settable option by which git-sync will not abort on a failed first git sync attempt, but would keep retrying.
This seems a bit counter-intuitive. I've found, however, that restarting a git-sync deployment that is seeing a transient error is less confusing than a git-sync that has reached a full 5 minute backoff delay and then gets fixed -- because if you are not careful you may miss the fact that you are seeing a backed-off git-sync.
In practice, we end up power-cycling git-sync anyways. At which point, we may as well keep git-sync up and running.
Long story short, PR upcoming.