Why the spike on link down for 1 hour

[gilson]

I monitor one remote frame relay router (Motorola VANGUARD 6455) regularly. It's traffic avarage is around 1 Mbit. The link has the 2 Mbits MAXIMUM bandwidth (not burstable) limit.

But yesterday, it went offline for about 1 hour (in fact, my link to it). After that, the graph drawn a spike around 10 Mbits.

See it at http://200.162.5.70/shot.gif.

Why this happened ? It looks snmpget returned the "acumulated" traffic.

I had, other time, a similar situation where the graph continued from it left but with a blank portion in the gap.

Any ideas ?

[raX]

Hmm I think you may have experienced your first SNMP counter overflow

Traffic is gathered using a COUNTER (an increasing number). When a device goes down suddenly, this counter gets reset. So the next time rrdtool goes to take the difference of the last two counter values; the result is extremely large.

One way to avoid this is set a limit for the data source so lets say numbers over 2 mbit would not show up. MRTG does some things to detect this behavior, but I've still had it happen to me.

I know it's not the best answer, but this is something that will have to be addressed in the future.

-Ian