[Nagiosplug-devel] comparisons of -w and -c
Andreas Ericsson
ae at op5.se
Thu May 5 13:44:58 CEST 2005
sean finney wrote:
> hi,
>
> On Thu, May 05, 2005 at 10:09:17PM +0200, Andreas Ericsson wrote:
>
>>Josh wrote:
>>
>>>The check_disk and check_swap plugins error if -w is *less* than -c - it
>>>should
>>>be the other way around. (this still is true in the most recent CVS)
>>>
>>
>>No it should not, since they report (and warn) on how much space (in
>>percent) that is free. It's a common misconception.
>
>
> and this behaviour has changed *at least* once in the past, if not more.
>
> honestly, for the sake of stability and consistancy i'd suggest leaving
> it as it is. it might be nice to have a flag that let you select what
> you'd like to check (% left/used and Gb/Mb left/used, like the -T in
> check_snmp_storage), but otherwise i think it's fine the way it is.
>
What's really needed is something like a general functionality to check
if one value is greater than the other and evaluate it intelligently,
like so;
/* eval thresholds in a somewhat clever manner */
inline int nplug_thresh_match(warn, crit, value) {
if(crit <= warn) {
if(value <= crit) return STATE_CRITICAL;
if(value <= warn) return STATE_WARNING;
}
if(crit >= warn) {
if(value >= crit) return STATE_CRITICAL;
if(value >= warn) return STATE_WARNING;
}
return STATE_OK;
}
This should ofcourse be called once for each threshold type (several
disks at once, fe).
Also, check_{disk,mem,swap} and friends should be made to accept a
prefix to the threshold arguments, such as %, MB et al. They should also
follow the path of least surprise for the end user.
I've never quite understood why there are a plethora of various
functions for exiting while reporting user errors when there are none
that actually work around user stupidity.
--
Andreas Ericsson andreas.ericsson at op5.se
OP5 AB www.op5.se
Lead Developer
More information about the Devel
mailing list