BigFix attempted to come up with a work around which starts here.
After reading about it, I chose not to test or implement it as my process was solid enough.
I’ve deployed 1809 using my same process and I’m willing to bet the issue is there. However, 1903, which I have not yet deployed with BigFix may not have the issue. I’m only guessing at this, but based on my manual install tests of 1903, the process has changed (hopefully for the better).
The change that I have seen is that much of the upgrade seems to take place during the execution of the setup.exe from the media (close to an hour). Then the first reboot goes to about 68%, followed by a second reboot which completes the installation. So that is about 2 reboots less than previous versions.
Based on that, I’m taking a leap of faith that the 1903 native fixlet when it is release will work well.