@39199a26 maybe "immutable" isn't the answer to everything. At least for systems where reboots are disruptive.

If your hypervisors run clustered at least sequential reboots won't hurt your payloads.
@2ea2dcbf