[rrd-users] Re: Hov the get max and average values for an interval

Bjorn Nordbo bn at nextra.com
Wed Jun 14 11:25:20 MEST 2000

First of all, thanks to those who helped me for prompt and useful

Dave Plonka wrote:
> On Fri, Jun 09, 2000 at 02:45:04PM +0200, Bjorn Nordbo wrote:
> > Where the numbers returned are the average values for the two data-
> > sources of dings.rrd (is the order of these fixed?) for the last
> > 24 hours.
> <snip>
> > Is this possible to do with rrdtool?
> You would use, believe it or not, "rrdgraph" to do this...
> I'm not quite clear, though, whether you mean that you want one average
> and max (of both DSes) or an average and max of each DS.  Let's assume
> you meant the latter and furthermore, lets assume you have both AVERAGE
> and MAX RRAs within the RRD.  You could do this:

That is correct.

>    rrdtool graph /dev/null \
>    'DEF:A=dings.rrd:ds0:AVERAGE' \
>    'DEF:B=dings.rrd:ds1:AVERAGE' \
>    'DEF:C=dings.rrd:ds0:MAX' \
>    'DEF:D=dings.rrd:ds1:MAX' \
>    'PRINT:A:AVERAGE:ds0 ave ave=%.2lf' \
>    'PRINT:B:AVERAGE:ds1 ave ave=%.2lf' \
>    'PRINT:C:MAX:ds0 max max=%.2lf' \
>    'PRINT:D:MAX:ds1 max max=%.2lf'

This works quite good, and it even seems like rrdtool is smart to
not draw an empty bitmap to /dev/null, so the performance penalty
should be very small if not non existant. Almost elegant I'd say. :)

> BTW, my RRGrapher front-end/utility (which you can find linked from
> Tobi's site and is here http://net.doit.wisc.edu/~plonka/RRGrapher/)
> produces an example command like this for you if you turn on its
> "[ ] Show statistics using this format:  [ %.2lf ]" feature under
> "Graph Options".

Thanks for your tip, it looks useful though I'm not yet quite sure
how I'm going to fit it into our statistics framework.

Best regards,

Bjørn Nordbø  -  IP Development  -  Nextra Norway

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://www.ee.ethz.ch/~slist/rrd-users

More information about the rrd-users mailing list