@ Martin: Thank you! Classical hasty reading mistake on my part then, I didn’t see that line about editing the file and skipped the step. 
I can’t entirely judge it yet, but for now I’m seeing a change for the positive.
Utilization spiked as to be expected right after reactivating the site once more, but it seems to be mellowing down. I’m still seeing spikes to 100% on all cores in htop if multiple system checks run at once, but they last notably shorter and don’t seem to be making a visible impact on the graph level anymore.
I’ll observe it some more to be certain since it has not long been long since the change yet, but it seems significantly improved. 
- it seems to be indeed stabilizing at a notably lower and stable level, except for the Check_MK Discovery, which still incurred a large CPU spike and many (but less than before) failures/timeouts.