[rrd-users] MAX values dependant on width of graph

[Simon Hobson]

Richard Burton wrote:

>I wonder if someone could explain this to me, either I am missing
>something or I have some errors in my scripts.
>
>It appears that when producing a graph the MAX values are not calculated
>on the actual absolute maximum value in the sampled data for the given
>time scale specified, but on the maximum value displayed on the graph
>after the data has been aggregated to fit into the required pixel width.
>As such generating the same graph twice but with different width
>settings produces 2 different MAX values?

Are you certain that both graphs are using the same dataset ? When 
changing the timespan of the graph, it's fairly easy to change which 
RRA is used to produce the graph which could alter the values 
returned.

Also, I assume you are storing an RRA for MAX and not taking "MAX of 
AVE" which is an easy mistake to make.