[rrd-users] Re: xport values and Cricket logs don't match up

[Alex van den Bogaerdt]

On Tue, Dec 05, 2006 at 08:21:17AM -0500, Goldblatt, Eric wrote:

> When I convert the server times (Eastern Standard Time) to unix
> timestamps, take the differences of the inOctets counters from the
> Cricket log, and divide by 900 to get bytes per second, I get the
> following:
> 
> Date  Time (E.S.T.) RoundedTime(EST) UnixTimestamp
> CricketLogInBytesPerSec       RRDxportInBytesPerSec
> 27-Nov-06     5:49:05 5:45:00 AM      1164624300
> 
> 27-Nov-06     6:04:07 6:00:00 AM      1164625200      5.7831777778E+02
> 5.8599484280E+02    
> 27-Nov-06     6:19:05 6:15:00 AM      1164626100      5.8053333333E+02
> 5.8036120522E+02    
> 27-Nov-06     6:34:06 6:30:00 AM      1164627000      4.1648555556E+02
> 4.6121499652E+02    
> 27-Nov-06     6:49:06 6:45:00 AM      1164627900      5.4108111111E+02
> 5.0687749074E+02    
> 27-Nov-06     7:04:07 7:00:00 AM      1164628800      4.9322111111E+02
> 5.0604922048E+02    
> 27-Nov-06     7:19:07 7:15:00 AM      1164629700      5.6567666667E+02
> 5.4554124742E+02    
> 27-Nov-06     7:34:06 7:30:00 AM      1164630600      5.2068666667E+02
> 5.3314918736E+02    
> 27-Nov-06     7:49:07 7:45:00 AM      1164631500      8.1644333333E+02
> 7.3489891666E+02    
> 27-Nov-06     8:04:08 8:00:00 AM      1164632400      7.3416333333E+02
> 7.5641274117E+02
> 
> I expected the values in the last two columns to agree almost exactly.
> What am I doing wrong?

You are rounding 6:19:05 to 6:15:00; this is 4 minutes 5 seconds, which
is 245 seconds.  245/900=27.2 percent.

You get 5.81e2 and 5.80e2, the difference is  0 percent.
You get 4.16e2 and 4.61e2, the difference is 10 percent.
You get 5.41e2 and 5.07e2, the difference is  6 percent.
[...]

Why do you consider 27 percent difference OK, but 10 percent too much ?

I suggest you brush up your skills on normalization (see my site) and
try a more accurate computation.

Without actually analysing the data, it looks quite similar to me. Each
time one series grows, so does the other. Each time one series drops,
so does the other.  I'd be surprised if the numbers don't match up after
truely normalizing them.

-- 
Alex van den Bogaerdt
http://www.vandenbogaerdt.nl/rrdtool/

--
Unsubscribe mailto:rrd-users-request at list.ee.ethz.ch?subject=unsubscribe
Help        mailto:rrd-users-request at list.ee.ethz.ch?subject=help
Archive     http://lists.ee.ethz.ch/rrd-users
WebAdmin    http://lists.ee.ethz.ch/lsg2.cgi