[rrd-users] omit invalid values

Alex van den Bogaerdt alex at vandenbogaerdt.nl
Fri Nov 21 16:47:56 CET 2008


----- Original Message ----- 
From: "Ulrich Stärk" <uli+lists.rrd at spielviel.de>
To: <rrd-users at lists.oetiker.ch>
Sent: Friday, November 21, 2008 2:14 PM
Subject: [rrd-users] omit invalid values


> Hi,
>
> I'm new to rrdtool so please be patient with me :-)
>
> I used ntop to populate some rrds and now want to generate custom graphs
> from those. Unfortunately some really unreasonable values got stored in
> my rrds which I neither want to be printed in my graphs nor be used for
> calculating the min/max/avg values.
>
> How can I tell rrdtool to just ignore values above some threshhold that
> I specify?


First you need to think about *when* you want to do so:

1: when providing rrdtool with input, thus "rrdtool update"
2: when creating graphs

In case of 1 you need to do some work, but this needs to be done only once:

You should have set appropriate minimum and maximum rates when you created 
the database. See http://oss.oetiker.ch/rrdtool/doc/rrdcreate.en.html and 
look for min and max.

But you already created one, so you need to "repair" it: see 
http://oss.oetiker.ch/rrdtool/doc/rrdtune.en.html on how to set these new 
min and max rates.

Use http://oss.oetiker.ch/rrdtool/doc/rrddump.en.html and 
http://oss.oetiker.ch/rrdtool/doc/rrdrestore.en.html to make a backup and 
restore it again, see 'range check' in restore.

After doing "rrdtool tune", rrdtool will ignore rates when they are outside 
the limits you set.  After going through "rrdtool dump" and "rrdtool 
restore --range-check" rates in your RRD should be within the limits you 
specified.


In case of 2 you need to do less work but your computer will need to do this 
every time you create a graph.  See 
http://oss.oetiker.ch/rrdtool/doc/rrdgraph_rpn.en.html and look for LIMIT 
(under comparing values).
This way you keep those faulty rates, you just ignore them every time you 
create a graph.


HTH
Alex 



More information about the rrd-users mailing list