for basic heartbeat I check processes (ps -fu oracle | grep _${ORACLE_SID}) and timestamps of controlfiles (should be newer than system clock - checkpoint_timeout).
for more extensive "health" monitoring you need to separate (&understand differences between) instance from database. instance health is indicated by things like cache hit ratios (db & share pool), # of client processes, etc. while database health is indicated by things like tablespace utilization, extents management of objects, chained/migrated rows (bad), etc.
of course those are just basic examples since there are bazillions of things you can monitor in Oracle and what YOU need to watch depends largely on your application/architecture.
one thing I will say is if you're in it for the long haul learn your data dictionary and v$ views! do NOT use OEM! I may be a dinosaur (at the ripe old age of 33 but I won't consider a candidate who gives an OEM based answer in a interview (though I will tolerate rman even though I still use my own shell scripts for backups.
Hi,
There are lot of ways by which you can do the health checkups of your database. You can calculate data hit ratio by querying the dictionary and make an analysis of your database usage. You can also check out the library hit ratio and the library cache access details, by which you will come to know about the access of your dictionary cache details.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.