docs: add documentation for job priority = 0 behavior

[jameshcorbett]

I couldn't find any documentation on the consequences of a job having priority zero. From a conversation with @grondo on slack:

Hm, I'm not seeing anything explicit:  In RFC 30:

 A value of 0 indicates that the job should be held. A held job is assigned the lowest possible priority, bypassing the job manager priority plugin. Although it remains in the queue while pending, a held job SHALL NOT be assigned resources.But that is regarding urgency, not priority. However, the minimum defined priority is 0  and the job-manager does explicitly set the priority to FLUX_JOB_PRIORITY_MIN when urgency is 0.

Maybe this should indeed be clarified somewhere.
I think the job-manager may avoid sending alloc requests for priority=0 jobs too, so the scheduler won't even see them. I'd have to dig further to be sure though