[rrd-users] Newcomer's RRD design dillemmas

Tobi Oetiker tobi at oetiker.ch
Tue Oct 16 20:12:15 CEST 2012

On 16.10.2012, at 13:06, Simon Hobson <linux at thehobsons.co.uk> wrote:

> Ryan Kubica wrote:
>> 'Exact values for an interval' and 'exact value at a time' are 
>> different things.
>> To store the exact value you want with no averaging just store it on 
>> the step-size interval boundary, rrdtool won't average those.  ie: 
>> binary data is no problem 0/1 if stored on the interval.
> However, if the OP wants to keep that data for min 5 months, at 1 
> minute resolution, then that's a lot of samples. If, as is hinted, 
> changes are infrequent, then that's a very inefficient way of storing 
> them - ie there are other tools that would be more appropriate for 
> that.

I agree ... but just for the sakes of argument, if he was certain that there never would be more than one change per hour, I think it would be possible to reconstruct the time of change from an hourly average, if the update was applied at the time of the change.


> RRD will certainly store his other (scalar) data though.
> -- 
> Simon Hobson
> Visit http://www.magpiesnestpublishing.co.uk/ for books by acclaimed
> author Gladys Hobson. Novels - poetry - short stories - ideal as
> Christmas stocking fillers. Some available as e-books.
> _______________________________________________
> rrd-users mailing list
> rrd-users at lists.oetiker.ch
> https://lists.oetiker.ch/cgi-bin/listinfo/rrd-users

More information about the rrd-users mailing list