The connection to the database cluster was restored, production app servers were rebooted, and the app works fine again. We're just making extra sure that the data still has integrity, but there should be no problem since it seems that the DB itself never went down, just the network connection between.