RE: SQL*Net message from client
- From: "Cary Millsap" <cary.millsap@xxxxxxxxxx>
- To: <oracle-l@xxxxxxxxxxxxx>
- Date: Thu, 3 Jun 2004 23:13:34 -0500
Indeed. In Hotsos Profiler v4, we already emit range (min/max) information
for call durations. In v5 (currently in development), each resource profile
row contains a histogram /picture/ that gives an instant visual indication
of the skew for a given timed event's durations. In our early testing, this
little picture has proven itself to be extremely effective in conveying the
information the analyst needs without having to break into the raw trace
data.
Cary Millsap
Hotsos Enterprises, Ltd.
http://www.hotsos.com
* Nullius in verba *
Upcoming events:
- Performance Diagnosis 101: 6/22 Pittsburgh, 7/20 Cleveland, 8/10 Boston
- SQL Optimization 101: 5/24 San Diego, 6/14 Chicago, 6/28 Denver
- Hotsos Symposium 2005: March 6-10 Dallas
- Visit www.hotsos.com for schedule details...
-----Original Message-----
From: oracle-l-bounce@xxxxxxxxxxxxx [mailto:oracle-l-bounce@xxxxxxxxxxxxx]
On Behalf Of Jonathan Lewis
Sent: Thursday, June 03, 2004 2:27 AM
To: oracle-l@xxxxxxxxxxxxx
Subject: Re: SQL*Net message from client
Cary,
That looks like an opportunity for your profiler -
add a mean, and standard deviation to each
wait time you record, then add in the ability to
drill down to a histogram showing distribution.
It's interesting to note that one of the enhancements
on wait information on 10g is the histogram output.
v$event_histogram (system level, but with snapshots
it can be informative, and it's definitely interesting
even when it doesn't necessarily help on a specific
problem)
(And 10g does have the convenient v$session_wait_history,
so you can get the last 10 waits with one query - that can
help the real-time search).
Regards
Jonathan Lewis
http://www.jlcomp.demon.co.uk
http://www.jlcomp.demon.co.uk/faq/ind_faq.html
The Co-operative Oracle Users' FAQ
http://www.jlcomp.demon.co.uk/seminar.html
Optimising Oracle Seminar - schedule updated May 1st
----- Original Message -----
From: "Cary Millsap" <cary.millsap@xxxxxxxxxx>
To: <oracle-l@xxxxxxxxxxxxx>
Sent: Thursday, June 03, 2004 4:37 AM
Subject: RE: SQL*Net message from client
One point of clarification that comes to me upon seeing my own response
here... Just adding a COUNT column to the existing output isn't going to
reveal the answer you need. You really have to know something about the
distribution of snmfc durations (are they uniformly distributed? Or
skewed?). You'll only be able to see this by looking at the raw trace data.
It is impossible to see in the V$ data unless you poll the V$ fixed views
very rapidly (on the order of 100s of times per second).
----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com
----------------------------------------------------------------
To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx
put 'unsubscribe' in the subject line.
--
Archives are at //www.freelists.org/archives/oracle-l/
FAQ is at //www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------
----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com
----------------------------------------------------------------
To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx
put 'unsubscribe' in the subject line.
--
Archives are at //www.freelists.org/archives/oracle-l/
FAQ is at //www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------
Other related posts:
