Very intriguing article from Gorjan Todorovski with Pythian about how we can start monitoring the health of our DB in some surprising ways. This really has my mind spinning about how to implement some monitoring for the Support group, even though the article was more geared towards a DBA, I think that this is something crucial for Support to be aware of. I'm thinking of a temp table, or several, so that we can record the information and have it available for our own historical purposes of tracking performance even when some of the tables Gorjan references (which were new to me, so yay!) get their information purged out by design or DB reboot.
Extending this idea further, I'd like to see what I can do to track not only a few metrics (obviously Gorjan picked just a couple of well known wait events to model his idea with) in the DB but go into the Exadata storage cells to gauge their performance as well! Additionally, I've never worked with the epsilon subquery so I've got a new tool that needs to be sharpened for my toolbox!
No comments:
Post a Comment