Re: WAS:A Challenge

Well, I am not going to give a solution here, but suggest to use one feature of OEM - so called Extensibility. You can have your own metrics collected based either on SQL, PL/SQL or scripts (plus perhaps something else). This is configured via XML files and takes a bit of time to figure out how but once you've done that - it's relatively easy.
For example, I created a simple new target for monitoring tuxedo applications - it had only one metric - status but adding new ones is a piece of cake.

Regards,
Alex

2005/11/14, Kevin Closson <kevinc_at_polyserve.com>:
> >>>
> >>>We painfully made 9i OEM work...but 10g OEM kicks A$$! It
> >>>really does.
> >>>Agent auto-connecting... Built in policy violations..
>
> that is a long list of exiting things OEM can do. But I still
> don't see a richness of cluster database drill down. Here's
> an example. Use OEM to answer this one:
>
> "I have an 8 node cluster, 3 databases (OLTP,DSS,DEV) with
> varying instance counts and some node overlap, I have a mix
> of OLTP and DSS. I need to get the average I/O latency for LGWR
> on a per-instance basis for the OLTP instances (database OLTP)
> and the aggregate of MB/s scanned with PQO only on the DSS instances
> (database DSS). "
>
> Ok, go...but if you can't use OEM to answer that one, I know
> how it can be answer with the execution of one single command.
> --
> http://www.freelists.org/webpage/oracle-l
>
>
>

--
Best regards,
Alex Gorbachev
--
http://www.freelists.org/webpage/oracle-l