13 Oct
2011
13 Oct
'11
4:14 a.m.
On 13/10/11 12:59, Adam Armstrong wrote:
It's virtually impossible to build such a set of rules though, which is why the problem is so great.
It's difficult to use ifSpeed to 'guess' the speed of the interface, as it's not always correctly reported, particularly on virtual interfaces.
People with 16x10G trunks would be very annoyed if we started returning NAN for measurements above 10G!
That I mean is not build rules to limit "big guys". I'm talking about not writing 0 value to rrd when snmp can't get data. Maybe I miss your point, but why in first place we save 0 value in rrdtool instead NA?