On Monday 09 September 2002 08:02, Xor wrote:
> Hello, Nick.
>
> You wrote Monday, September 09, 2002, 9:22:10 AM:
>
> N> Hi,
> N> Isn't the usual way of collecting stats of this kind, you get
> the N> value when you start monitoring and then the value when you want to
> stop N> and take the difference. This way if you want measurements of
> different N> periods its only the programs doing the measureing that change
> not the N> device producing the readings, and you can have multiple readers
> of N> different periods.
>
> ... and what about overlimit counters ? Example: counters was max
> values two (or more) time every some times ? I will not argue... I
> understand that can to do read different more often, but reset
> counters is more suitable %)
That's not such an issue if you read the counters fast enough. If you read
the counters each 5 minutes, you know quite accurate when they are resetted
and you will not loose that much data.
Stef
--
stef.coene@docum.org
"Using Linux as bandwidth manager"
http://www.docum.org/
#lartc @ irc.oftc.net
_______________________________________________
LARTC mailing list / LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
