(...)
> > I think the standard perfdata should store as much info
> > as possible.
> > The "convert to a graph tool" can make the decision to=20
> > reduce down to integers.
I agree. Imagine responses that are usually well below 0.5, like ping =
responses in milliseconds. Wouldn't we have to rewrite the ping check =
(or some filter) to multiply the values by 10, 100 or 1000 before =
entering them in the integer-only RRDs (and then convert it back before =
displaying, because nobody wants to see ping response averages in micro- =
or nanoseconds)?
Karl DeBisschop wrote:
(...)
> Maybe the best answer is to use ints in most cases - leaving=20
> the basics of the current perfdata macro as they are. But to=20
> add a perfdataf function that makes floats for output.
If there's not some significant performance or disk space gain, I'd say =
go for floats everywhere, as that seems like a simpler solution than =
splitting up in two functions.
(...)
> So lets work it out in the=20
> next week or so, and hopefully the result will not entail big=20
> changes relative to 1.4.0alpha1
Godspeed!

This post was automatically imported from historical nagios-devel mailing list archives
Original poster: bjornar.bjorgum.larsen@ementor.no