FreeLists / oracle-l / Re: "Performance Counters"

[Job Miller <jobmiller@xxxxxxxxx>]

Niall mentioned that EM/GC measures this.  A large amount of work has been done 
to expand Service Level Management in Grid Control.  I think it is a big 
unknown by DBAs though.
   
  You define the transactions of interest typically web transactions by using 
an IE browser launched from within GC and you record the interaction with the 
application for a typical transaction from the end-user perspective. GC agents 
functioning as "beacons" continually replay it measuring its performance 
against your expectations or at least baselining what that level is.  
   
  If the application isn't achieving those metrics, than you dig into the 
Oracle equivalent diagnostic data to see why.   If the DBA tunes something, 
than it should show up in the numbers for the performance of that service in a 
easily quantifiable metric.  (response time)
   
  Job
   
   
  Niall Litchfield <niall.litchfield@xxxxxxxxx> wrote:
  you are going to have fun! In an ideal world I'd be looking at defined
transactions and setting targets for elapsed time EM can measure this.
On the other hand what they've given you are technical efficiency
measures, they've a long way to go.

On 4/27/07, Jeffery Thomas wrote:
> We have a SQL Server team going to manage a new vendor-built application on
> 10g RAC (Solaris).
> They have zero experience with Oracle. They are asking the Oracle team
> for 'performance counters'
> similar to what they use with SQL Server so that they can identify
> "satisfactory performance criteria" to be
> used as non-functional requirements for the vendor building the application.
>
> The ones identified for SQL Server were:
>
> SQL Server: SQL Statistics: SQL Compilations/Sec counter must not exceed 100
> per second over a 2 minute period.
> Average Disk Queue Length, averaged over a 2 minute period must not exceed
> 2.5
> SQL Server: Buffer Cache Hit Ratio should exceed 90% measured over a 2
> minute period.
> The CPU usage must not exceed 35% averaged over a 2 minute period.
> The SQLServer:LockWaitTime must not exceed 2ms measured over a 2 minute
> period
>
> If you've got a vendor building an application for you -- what kind of
> benchmarks -- if any -- actually make sense with
> respect to performance criteria that the vendor must meet?
>
> Thanks,
> Jeff
>


-- 
Niall Litchfield
Oracle DBA
http://www.orawin.info
--
http://www.freelists.org/webpage/oracle-l




       
---------------------------------
Ahhh...imagining that irresistible "new car" smell?
 Check outnew cars at Yahoo! Autos.