CEO & CTODuring a 5-6 minute period managers in some races may have lost connection to the race and some races may have restarted entirely as they were restored. We are investigating the cause now. The service seems to have recovered on its own, as it is programmed to do, but as a precaution we have immediately instructed all subsequent races to run on other servers which were reliable during the same period.
UPDATE: Logs indicate that this was some kind of network issue, likely a third-party problem (datacenter for example, not an issue with our service). We will restore the servers to operation overnight and monitor the situation. Apologies for any inconvenience this has caused, particularly to any leagues that had run their race to completion and then had it restart. I appreciate how frustrating this can be, but there is little we could do about this.
UPDATE 2: It seems quite probable that a maintenance window being carried out by one of our third-party server providers is the culprit. They warn that bad network latency may be an issue until 5am GMT on March 22nd.
To request that a race be re-run, please get the host of the league to post in this thread.