Hi!

I remember submitting a patch to Tobi that defines a CDEF called AVG, that does ignore NULL values. I now did check if this is in the SVN branch and I can confirm this is included.
Maybe this is what you can use...

Martin

P.s: I remember having used a very complex CDEF to work around this for just 4 values and the patch to rrdtool was much shorter ;)

Simon Hobson wrote:
George Barnett wrote:

I'm aggregating boxes into a graph to draw a cluster and I'm getting
null values.

Is there a maximum number of datasources in 1 graph?

Is there a maximum size of a cdef?

I've found that if I do this with only a few DS's, it works fine.  Maybe
somebody has tried this before?

<snip>

"CDEF:gclient_requests=1aggavgclient_requests,2aggavgclient_requests,+,3aggavgclient_requests,+,4aggavgclient_requests,+,5aggavgclient_requests,+,6aggavgclient_requests,+,7aggavgclient_requests,+,8aggavgclient_requests,+,9aggavgclient_requests,+,10aggavgclient_requests,+,11aggavgclient_requests,+,12aggavgclient_requests,+,13aggavgclient_requests,+,14aggavgclient_requests,+,15aggavgclient_requests,+,16aggavgclient_requests,+,17aggavgclient_requests,+"\

Are you aware that if ANY of the values is unknown them the result is unknown ? So for example, if you were not collecting samples for one system for a while, then you would have a gap in the graph.

You have to do a test for unknown values and replace them with zeros before adding them all together. Have a look back through the archives, it's come up before.

_______________________________________________
rrd-users mailing list
[email protected]
https://lists.oetiker.ch/cgi-bin/listinfo/rrd-users

_______________________________________________
rrd-users mailing list
[email protected]
https://lists.oetiker.ch/cgi-bin/listinfo/rrd-users

Reply via email to