> 
> DEF:vname=rrd:ds-name:CF[:seconds]
> 

If possible, it should also be possible to use a resolution that
is too high for the graph.

Suppose I have samples every 10 seconds, and would like to use
these samples to calculate the maximum value in one period,
it would not be good to first average the values (as it does now)

Example: In my previous mail I described the daily overview. It
graphs 432 pixels, 20 samples per pixel, for a total of 24 hours.
The 20 samples are averaged first and only then the maximum value
is calculated.
The 2-hour periods are built from 720 samples and since it does not
graph it is not averaged first (at least, I think that it works this
way ?). The maxima are therefore higher than in the graph.

This could be solved by adding an extra rra (containing maxima and
with a lower resolution) but I already have the data available ...

Regards,
Alex 

--
* To unsubscribe from the rrd-developers mailing list, send a message with the
  subject: unsubscribe to rrd-developers-request@list.ee.ethz.ch