Document how sqlalchemy handles transaction aborts or the best practice to handle transaction aborts (for the readwrite transactions)

Could you please help document the best practice to handle transaction aborts and the errors like “The transaction was aborted and could not be retried due to a concurrent modification.” There were multiple support cases asking about this.

Thanks!

We are about to merge this: #243
Thanks for reaching out!