One of our servers experienced a DNS failure. Our high-availability setup re-routed its IP to another server, so there was no extended downtime, but a small number of users may have experienced a few request errors during a short period of time.