How do you know when something is wrong with your DB?

When something abnormal happens, but not really crucial, how do you know it happened? Say, a datafile suddenly becomes corrupted, or you lose a redo log file, or some other act of supreme (and unfriendly) powers that does not cause Oracle to shutdown.

One way is, of course, all the phone calls from users who cannot use the DB anymore. But I am looking for something more proactive. For example, a cron job (on Unix) that will periodically scan the alert log file, look for errors and email them to me.

Any other ideas/suggestions?

Thanks,

-- 


[ Simon Goland       B-)>     sg_at_mda.ca ]

[   Without action there is no change   ]