My RHEL Machine is a Build server and always gets shutdown / rebooted at around 5:30 every evening. No idea whats going on as developers were unable to connect to the machine since three days. When I am visiting the physical server probably next morning (since last 3 days)its doesn't show anything and all i reboot the machine and it comes up without any issue.Any idea how to troubleshoot?

Any kind of scheduling? Any logs?

"Around" doesn't seem too exact to me, so I'd guess it's a hardware problem overall if it's a bit hot around there.

Second, there's always the chance of an intrusion. Shutting down network services and using rkhunter, chkrootkit and clamav is your best bet there. Of course that will only be possible if you are physically there.

Third, be vigilant, specially about temperature. When things start going off at random times for no reason that's as good reason as any other.

Last but not least, turn off crond and any other custom init script.

Cant it be checked through /var/log/messages?

It sure should give some hint in /var/log/messages

That entirely depends on how your box is configured. More particularly, the kernel will log something there. The rest is up to the system logger, and that can be configured to log most things or nothing at all.

Also, note that in the particular case of an intrusion there's a very high chance that the traces have been erased or modified to hide anything that could lead you to your attacker. Anyone that's smart enough to break into your system is probably not that dumb not to take that preventive measure.

I already gave you some hints on what to do. You should really be checking them.

Do check that something bizarre isn't happening physically. I did once hear a story of a hospital that had bad survival figures on a Friday, particularly. It turned out that on a Friday evening, the cleaner came round with a floor cleaner. Needing somewhere to plug it in, she had to unplug something. Now, she didn't know that it was a life support machine that she was unplugging...

In your case, you can think that things could be worse. I don't know what, in general, the hospital thought...

That's definitely been known to happen in offices; had it happen to me once.
What I'd do is sit at the machine at 5pm onwards and run

tail -f /var/log/messages

and

top

in separate xterms. The log files you've got may give you a clue, but simply sitting there is probably going to be quicker in the long run.
Possibly also

tail -f /var/log/audit/audit.log

as well.