How to beat irregular counter reset

[tata_tulen]

Hi all,

I'm trying to create templates for WatchGuard IPSec Tunnels Traffic and I'm facing the following problem: each tunnel has its time and length maximum when the tunnel is re-established, during suach a process the counters are reset. Therefore I get extreme peaks in graphs as RRD process them as "overflow"...

I've tried set the data length limit of tunnel as maximum value in graph, but it doesn't work.

Is there any elegant way to beat this behavior? Each tunnel has timestamp with datetime (encoded in hexa value that I hasn't decode so far) - is there any way to force RRD not to use "overflow" in case of reset the tunnel?

Thanks for any hint!

[stefanbrudny]

Thanks for any hint!

Any hint you say... not exactly prevention, but I use spike kill plugin for removing spikes. It is a little limited, but works most of the time.

I'm also looking for such restriction.