[rrd-users] Usual RRD question on AVERGAE, MAX and consolidation :)

Johan Elmerfjord jelmerfj at adobe.com
Thu Dec 19 16:54:17 CET 2013

Hi James,

So in your graphs you are now plotting the MAX-values I assume.
When more consolidation occurs (in the smaller graph), you will get a higher average as well as it may look for the highest value out of 4 data-points,
while it's in the wider graph creates the average out of  say the max of 2 datapoints.

I would strongly suggest sticking to the average-numbers (of the AVERAGE database)- as the max can be very confusing, and may be way of reality.

You can however pull the Max-value from the MAX-database and display that under the graph.
And if you don't consolidate the data as time goes, you don't really need to store a MAX- database either, as you can override how data is consolidated.
And can get your max of max as:

Good luck!


On Thu, 2013-12-19 at 15:14 +0000, James Bensley wrote:

A user has emailed me off list to inform me that data is consolidated
when you are drawing it into a graph, so this is like a second stage
of consolidation. After storing the data and possibly consolidating
it, the data is consolidated again if the graph you are producing
doesn't have enough pixels to display the data you are attempting to
pull from the rra's into a graph.

That sounds logical and strait forward to me, so I have tested this.
I have generated a graph using rrd tools on the CLI and uploaded it here;

I have generated another one but specified double the width and height
for graph size, and uploaded it here;

The min and max values are the pretty much the same but the average has changed.

So this looks like my question is answered unless anyone else any any
input on this?

Many thanks,

rrd-users mailing list
rrd-users at lists.oetiker.ch<mailto:rrd-users at lists.oetiker.ch>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.oetiker.ch/pipermail/rrd-users/attachments/20131219/0c27f532/attachment.htm 

More information about the rrd-users mailing list