Preemptive crash correction via ARD and/or LOM?

Question

I've been having some very strange crashes that no one can seem to figure out. Problem already posted here:
http://discussions.apple.com/thread.jspa?messageID=11366714&#11366714
Had a couple consultants already out and they can't find anything and nothing ever stands out in the Console logs. Gonna try a kernel dump probably next week when my users are all away (summer break at the university it's at). There's still staff over the summer that need to log in for events that happen over the summer and I don't have to be on campus for those times. The crashes usually happen overnight and I won't know about it until the person calls me--often only minutes away from needing the system online (and it takes 10-20 minutes for my current technique of using LogMeIn and navigating through Server Monitor and initiating a restart that way).
ANYWAY, in the mean time, is there some way I can be notified when it's down before a user has to call me? I'm sure the server can't let me know directly when it crashes, but I've got an ARD machine constantly connected that would know when the server goes offline. Is there some way it could email me when such an event happens? If not through ARD, is there some other program that can continually monitor the server and send me alerts?
Related, is there a program that can remotely interface directly with the LOM module? Ideally from my iPod. I guess theoretically I could use Server Monitor, but my mac network is sandboxed inside the main campus' PC network and the PC guys have been running me around for the past 10 months about getting me my own external IP address to connect to my internal router (that's connected to the main network to provide my network with internet access). However, an ideal solution would be something directly from my iPod. I've got a couple other handy programs I like (Server Admin and iStat), but neither of them can talk to the LOM (so not worth burning down the PC IT director's door to make sure they can get through once I'm off campus).
Thanks.

Answer 1

Again, it's great having you on speed-dial here, Antonio. I've been thinking of external faults lately, myself, but that's SO troublesome to pin down. It's also so intermittent that a "solution" may not be readily apparent (I've had it crash twice in two days and go almost two weeks without incident). I'd been looking at Splunk to help me catch anomalies across my entire network, but being so new to that level of reporting, I haven't really made heads or tails out of that, yet.
I stopped using Time Machine for backups a while ago (I had suspicions the hourly backup was interfering with consistent video captures so stopped using it to be safe). The UPS is brand new (purchased at the time of the server last August), so I'm hoping it's not that, but will try taking it out of the loop as well. Most of my switches/routers are brand new, but I do have an old switch (10/100 that I'm using for some old eMacs because I ran out of ports on my new gigabit switch) and an old wireless router I'm using in a rarely used area.
Failing that, I'm screwed if it's a cabling issue. Most of it's a new install, but I honestly don't know how up to par the entire network cabling is--there's a few higher ups that like to cut corners to "save money", so there's a chance for anything.
At least we're in summer break now so I can take it down to just a complete sandbox with direct connects to a minimal number of clients and no outside access to give it a "clean" a start as possible and slowly work back up...
And cpragman, I saw the SSH iPod stuff, but the server wasn't responding to SSH from a local client machine when crashed, so I didn't pursue it further. It just blows me away that I've seen dozens of apps for PC servers and only 2 or 3 for Mac products...you'd think Apple would be all over interconnectivity of all their hardware.