[rrd-users] Newby questions

Tobias Oetiker tobi at oetiker.ch
Tue Oct 30 07:35:09 CET 2012

Today Pablo Chacin wrote:

> On 10/29/2012 11:57 PM, Tobias Oetiker wrote:
> > you can track the amount of unknown time too ... but you can use
> > the TOTAL function will ignore unknonwn values .
> Didn?t know this. Could you please point-me to

you can use a CDEF to turn UNKNOWN to 0 and all the rest of the
entries to 1 calculate the AVERAGE. This will tell you how much
'UNKNOWN' is contained in your data stream.

> > otoh, why would you want to count UNKNOWN as ZERO ? isn't more
> > corect to asume that the values during the time where you have no
> > information were the same as in over the rest of the interval ?

> In my case, no. I may have gaps in the collection which, if i understand
> correctly will be traduced as unknown.

yes ... and thus, if you build a sum over all the entries, the most
correct solution is that the SUM is also unknown. RRDtool has the
option of ignoring the unknown entries using the TOTAL function
(which will cause them to essentially be treated as 0).

or if you take the average and and multiply with the total
interval, you assume the unknwon samples to be equal to the



Tobi Oetiker, OETIKER+PARTNER AG, Aarweg 15 CH-4600 Olten, Switzerland
http://it.oetiker.ch tobi at oetiker.ch ++41 62 775 9902 / sb: -9900

More information about the rrd-users mailing list