I know this is more of a linux-adminish type question, but I was hoping to still find some helpful clues :)
We are running our JSS on Ubunutu Server hosted by corporate offsite. But for some reason (probably a policy I created or something caching mavericks), it's pegging the MPLS.
I've heard ntop is a good way to go, but wanted to know what others were using.
For better or worse, I use graphdat
... that said, that's an interesting symptom, or description of a symptom. MPLS isn't a device or server. It's a transport type. Also, when these type of questions popup, they almost always end up with "Just what the hell are your network engineers doing" sort of response. Just an observation. Then again, I am not an expert and don't have any real experience with MPLS.
I once watched a major Ivy League University harass their Casper admin about their servers multicast traffic, when it wasn't pushing any. It turned out that they just blamed her for the campuses Apple products sending mDNS requests as they are apt to do.
Have you checked a clients Jamf.log to see what (if anything) hit the machine during the time frame your MPLS utilization peaked.
That at least could give you an indication of what may be causing the issue.