How do you configure your OEM events?

From: Dave <shoad316_at_hotmail.com>
Date: 7 Apr 2003 10:58:26 -0700
Message-ID: <78cf0572.0304070958.5070cab7_at_posting.google.com>



Hey all,

  I'm in the process of setting up OEM to monitor our environment. I have used OEM in the past and have always wondered how other people setup their events. I'm debating between two options:

Option 1 - Group events based on severity or focus. -=-=-=-=

Some examples
- Have an event that monitors space issues. (Tablespace, dump
destinations, archive disk, etc.)
- Database Up/Down

  • Listener Up/Down
  • Node Up/Down
  • Critical events (corrupt blocks, archiver hung.. events that require immediate attention)
  • Performance events (buffer cache, library cache, etc.. anything performance related..) . . . . you get the idea.

Option 2
-=-=-=-= - Group by Database

  • Put all of the events I want to monitor for a particular database into one event.

In the past I have used option 1. The biggest issues I have with this path is that if I want to add a new node/database to be monitored I have to deregister all of the events, add the node and resubmit them. Sometimes the Intelligent Agent gets out of sync and you have to clean it up.

The other con is that some databases may not need to monitor all of the tests in a particular event or on different schedules. Although this is rare, it can be annoying.

I guess, another option (Option 3) would be to create template events and use CREATE LIKE for each database I want to monitor. This allows me to customize it and not have to deregister events and reapply them.   The only reason i'm not fond of this one is that i'll have a gazillion events registered. While not really an issue for some reason i've tried to avoid this in the past.

I read Oracles whitepaper on OEM events and I was hoping it would go into more detail on how they setup their own environment. It sounds like the template idea is the best way to go.

Would anyone else like to share how they have configured OEM to monitor their environment? Does anyone know how to easily distinguish between 9i and 8i (and less) events without trying to register them or read up on each one?

Thanks
Dave Received on Mon Apr 07 2003 - 19:58:26 CEST

Original text of this message