This is how I typically design glass roof railings. Since they aren't explicitly addressed in the code, I've compared a few approaches and settled on parapets (with the rail functioning as an extension of the parapet height if it's attached to a parapet) as both the most similar condition that is explicitly addressed, and the most conservative of the potential options. I have seen one design where the glazing shoe for a rail was vertically offset from the roof surface and laterally offset from the parapet face by long (1'+) outriggers and I could see the argument for that case not needing roof suction to be considered, but I really rarely see detailing like that.
Regarding the joints, for what I work on they're usually ≤1/2" on a 5' mod (<1% of the length) with the gaps between glazing shoe segments usually being much smaller. I'm not aware of a way to justify a reduction in wind load or that roof suction won't be present on the back surface because of these gaps, but I'd love to learn if you are!
This structuremag article might also be helpful https://www.structuremag.org/article/frequently-asked-wind-questions/
------------------------------
Renn Henry, PE
Staff Engineer
------------------------------