Why not 3? It is about how paranoid you want to be about uptime. Uptime and production grade are two different things. There is certain uptime guarantees that a cloud provider will give. Production grade is durability of the product itself.
Algorithms that tolerate Byzantine failures generally require t<n/3, i.e. at least 4 nodes to tolerate 1 byzantine fault. This is in fact required for primitives as simple as reliable broadcast. Systems that use leader election such as Raft are not Byzantine tolerant. They tolerate t<n/2, meaning that 3 nodes does indeed buy tolerance to 1 crash, but it doesn't support anything else than crashes.