Re: Management Reports

This type of request is entirely too common. The problem is that measuring the "performance" of an Oracle system isn't like measuring the performance of a car in the Indy 500 by timing laps, but that is typically the perception of those asking. At my last employer, we developed application-specific performance monitors - how many orders pending at each stage of the process (approval, submitted, filled, etc.), average processing time (overall and per phase), etc. It was much more useful for identifying problems. Of course this added non-trivial load to the system.

The best such "single measure" chart I've ever seen was part of the Precise (nee Savant) Diagnostic Center. It was the response time of the periodic poll to gather information. Other products may have something similar.

What are you trying to measure? If it is generic response time, that approach might work. If you don't have anything already, you could write something that pulls some statistics periodically (not too often) and time it on each run, graphing how long it takes. Beware of the Heisenberg uncertainty principle though. The act of detection will likely influence the results. Performance problem may appear and disappear between polls - not showing up. Increasing the frequency of the poll will further denigrate the overall performance. This is a Pandora's box of potential issues. Also beware that an application-independent generic "performance graph" will not always reflect what the users see (e.g. due to application-specific issues - locking, etc.). Generic response time may not be what they want - and probably isn't. One problem with this approach is that often the generic response time is best when the application performance is at its worst! Why? Something else in the stack goes into a catatonic state so that the instance/database is freed up to service the response time query faster. I've seen it happen many times. [Actually, this can be a plus. Scenario: Them: "Quick! Check the database! It seems slow!" Me: "No, its actually running much faster than normal. It seems that the app isn't submitting any/much work."]

Another possible alternative is the "Etch-a-Sketch" approach. Give them whatever kind of nonsense appeases them to get them off your back so you can do useful work ;-) Of course, there are potential "issues" with this as well!

Don Granaman
[certifiable OraSaurus]

• Original Message ----- To: "Multiple recipients of list ORACLE-L" <ORACLE-L_at_fatcity.com> Sent: Thursday, June 13, 2002 10:33 AM

> Good Morning Everyone!
>
> My management wants a chart that shows the performance of the
> database. If this was your boss, what would you show them?
>
> Thanks,
> Mike
>
> P.S. This is a repeat e-mail. I never saw my other one hit the
> list.

-- 
Please see the official ORACLE-L FAQ: http://www.orafaq.com
-- 
Author: Don Granaman
  INET: granaman_at_cox.net

Fat City Network Services    -- (858) 538-5051  FAX: (858) 538-5051
San Diego, California        -- Public Internet access / Mailing Lists
--------------------------------------------------------------------
To REMOVE yourself from this mailing list, send an E-Mail message
to: ListGuru_at_fatcity.com (note EXACT spelling of 'ListGuru') and in
the message BODY, include a line containing: UNSUB ORACLE-L
(or the name of mailing list you want to be removed from).  You may
also send the HELP command for other information (like subscribing).