[rrd-users] trying to understand the relationship between source data, what's in rrd and what gets plotted

Alex van den Bogaerdt alex at ergens.op.het.net
Sat Jul 21 14:19:44 CEST 2007

On Sat, Jul 21, 2007 at 08:03:34AM -0400, Mark Seger wrote:

> yes, I agree completely about the ability about rrd's strength being to 
> be able to store data long term.  but if take a sample every 5 seconds 
> but then can't look at it at that granularity what's the point?  I did 
> see the suggestions about using multiple lines/averages for each 
> variable, and if you only have a few I suppose that could work too, but 
> I'm trying to deal with hundreds.  perhaps the best compromise is to use 
> gnuplot for fine grained plotting and using rrd for archiving.

Please let me summarize in my own words:

If you display 4000 rates on a 400-pixel wide graph, you want to
see 10 dots per column, not a line or an area connecting an
minimum, average, maximum or last.


If so:

This would require programming (in rrd_graph.c and friends).

In my example of 10 dots per column, how do you suggest RRDtool graph
should show these rates: 1 1 1 3 4 7 7 7 7  ?

Should it show 6 dots of equal size,
should it show 2 large dots and two small dots,
or ... ?

If my summary is not correct, then what do you mean?

Alex van den Bogaerdt

More information about the rrd-users mailing list