[Date Prev]   [Date Next] [Thread Prev]   [Thread Next] [Date Index]   [Thread Index]

 

     [snips-users] about NetIErr & NetOErr

I have taken a look at a couple of the NetIErr and NetOErr calculations.

They are expressed as a PERCENT of the total packets over a very short sample interval.

Does anyone else think this is not the best way of reporting a devastatingly important but tiny number? By the time you got to seeing more than 1% of packets over such a short period, performance on your device is smoking wreckage.

I would like to propose that these be reported as counters, if at all possible, deltas since the last time hostmon ran (that is, two or three minutes.) If you have seen any Ethernet errors at ALL in the last 2 or 3 minutes, you have a problem. Then maybe look at delta-time and report the number as something like 100x errors per minute (if it was reported as errors per minute it could easily be < 1.)

Is it realistic to remember the values from the last iteration? Obviously you would have to not report the first run through after hostmon-osclient started up.
Daniel MacKay
Canadian Information Processing Society, Information Services Professional
Halifax, Nova Scotia, Canada, +1 902.499.0488


Zyrion Traverse Network Monitoring & Network Management Software