I am running a virtual private server (Debian/Linux running apache+tomcat+jruby) and from time to time it becomes unreachable (I am not even able to ping it). I blame this on our service provider. When this happens the server has to be rebooted via the provider's management console and everything is back to normal. I have tried going through the logs in var/log but found nothing extraordinary there. What are the next steps in investigation such incidents?
It might have been helpful if you'd mentioned what you were looking for in the logs and failed to find.
There's probably not a lot you could do here. But you might try setting up a basic internal watchdog - using cron (or preferably a daemon) to fire and log events at regular, short intervals - then check if it ran continiously until the reboot.
And increase logging.
Presumably you are paying the service provider - you should be asking them to provide details of their investigation
These are the first questions I would want to answer in order to troubleshoot the problem:
Do you have a KVM you could use to try to establish a remote console session? In any case, you will want to see what is going on from inside the server at the time of failure.
When the logs do not indicate errors and when you do not have the ability to establish a remote console setting, you could settle for a