Actually receiving zaps with very high uptime is fairly complicated to implement well -- requires multi-geo redundancy and stuff like that...