In distributed systems getting to a failing state is exactly the same as getting to a succeeding state, both have to make CAP consistency/availability trade of

In distributed systems getting to a failing state is exactly the same as getting to a succeeding state, both have to make CAP consistency/availability trade offs. So in many cases not failing is just easier.