[rrd-users] Average problem

Simon Hobson linux at thehobsons.co.uk
Mon Nov 19 16:20:08 CET 2007

Gwenael Lahay wrote:

>Indeed, i put a wrong command in my previous mail. The frequency is 
>not correct but the other parameters is good.
>Concerning my problem, does it mean that i should use a very hight 
>value for xff (like 0.9 or 0.99) ?

You need to consider your requirements.

If it is important to get a value returned at 10:30 when you only put 
in the one value, then yes, you will have to use a large xff value 
(at least 0.97).

However, that means you will still get a value returned at other 
times when 29 of the 30 samples are missing - or 25 are missing, or 
10, or ...

So you need to consider whether it's more important to have the value 
- or to show that input data was missing.

More information about the rrd-users mailing list