Taken from the readme:
> the behavior of a cluster if it fails while such a manually-controlled transaction is not yet defined
Do you envision manually-controlled transactions working seamlessly one day? If not, have you thought about what guarantees are feasible/you hope to provide?