APM should naturally arise from the breadth and pace of decisionmaking opportunities. Any action that does not amount to making a choice between at least two relevant alternatives tends to feel unnecessary and superfluous, and imo is a cheap way to make the game "harder".
As an easy example in early sc2 betas (iirc, don't quote me) many players complained that the game was too easy. The developers ended up adding some additional mechanics to artificially raise the skillcap. One of these mechanics was larva inject. Zerg players need to periodically use the "inject larva" ability with their queens in order to spawn larva that can be used to construct units. Without larva you can't make units. There are some edge cases where players skip injecting larva to save queen energy, but in the vast majority of cases injecting does not include a choice from the player and is just a mechanical task that players have to perform periodically. Furthermore the few edge cases where you'd skip injecting are only relevant for high level players and as such have no effect on how an averge player perceives the game.
In my view, pausing massfabs is an entirely mechanical task as there's no situation where you wouldn't want to pause your massfabs when running low on energy. That's why automating massfab pause is fine (as long as everyone has access to it). For the same reason automating pausing buildpower is not fine -- there's compelling decisionmaking in which engis to pause etc.