An AppEngine service to do task scheduling on highly hetegeneous fleets at medium (10000s) scale. It is focused to survive network with low reliability, high latency while still having low bot maintenance, and no server maintenance at all since it's running on AppEngine.