Earlier quoted context omitted.
You can buy a completely dev-usable Mac for under $1000. (My CI machine is a Mac mini I bought used for $350 and dropped an SSD and an extra stick of RAM into.) Its "crappy dev tools" are free . This is some silly stuff right here.
How reliable is your CI machine? I don't have the log entries anymore but in the past I've seen Mac Minis that were set up as Jenkins nodes throttle hard, leading to variable job results. The Minis were running VirtualBox VMs which would arbitrarily report long running test failures. I remember thinking perhaps it was VirtualBox specific behaviour until I search for "mac mini cpu throttle" and found similar results.
(It also runs a couple other small things for the house.)