[rrd-users] Re: Excluding UNKN or NAN from CDEF

Eric Brander Eric_Mailing_List at rednarb.com
Mon Sep 27 16:59:19 MEST 2004

Stephan Harren wrote:

> Hi Eric,
> I think, it might be good for you to read the tutorials and manuals 
> about rrdtool. But anyways, what do you want to achieve? Showing 
> everything allright when there are NaN's? It is not very nice to say: 
> "Okay, if I don't know what happened, probably everything was allright.".

Well, not really. I'm still graphing ups and downs (1's and 0's 
respectively) but I only want the service level line to report the 
average of everything EXCEPT unknowns. I don't want the SLA to be wrong 
because the polling PC missed a poll. Therefor I only want actual data 
to be graphed and unknowns ignored.

> But ok, there is one line that defines Cleaned:
> CDEF:Cleaned=Response,UN,0.0,Response,IF
> This line sets all NaN to 0.0, insert whatever you want instead of the 0.0.

Good suggestion but its result is not quite what I was looking for. 
This makes all unknowns = 0.0 (or in my case I tried 1.0 - which made 
the graphs look far more successful than they should be) Again, the only 
acceptable result would be to eliminate unknowns entirely.

> Cheers,
> Stephan

I'll go run through the tutorial pages again. It has been a while and 
though I felt comfortable with RRDTool afterward, this seemingly more 
complicated graph building syntax has be a little bewildered.


Eric Brander
All questions to life's problems are easy, if you know the answer.

Unsubscribe mailto:rrd-users-request at list.ee.ethz.ch?subject=unsubscribe
Help        mailto:rrd-users-request at list.ee.ethz.ch?subject=help
Archive     http://www.ee.ethz.ch/~slist/rrd-users
WebAdmin    http://www.ee.ethz.ch/~slist/lsg2.cgi

More information about the rrd-users mailing list