Yes, the same considerations for each HA pair apply. You can either use “system node halt -inhibit-takeover” for each node or globally disable takeover using “storage failover modify”.
I guess cluster is never expected to be switched off completely, so nobody thought about documenting it … ☺