How about an EXOS TCL script scheduled, so the XMC would log in to the switch, perform ping or show fdb or anything else and by parsing the output ($CLI.OUT variable) you could see what's going on...?
From EXOS side you could play with UPM scripts and EMS-based event trigger so once a port is down (or else) you would take care.
Or you can do that without any need for proactive action - an edge switch should syslog its events, and with XMC alarms you can wait for appropriate event showing port X is down so the alarm goes up.