Problems graphing interface errors

[jon3k]

I'm trying to graph the errors on a 3500 series cisco switch, and I end up with the attached graph.

Now, maybe I'm reading it wrong, but, unless the counters are cleared - the number of errors should never go down .. right? Also, how is the maximum less than the average? Something doesn't add up here (literally )

Any feedback appreciated

[gandalf]

Now, maybe I'm reading it wrong, but, unless the counters are cleared - the number of errors should never go down .. right?

This is how rrdtool works with COUNTERs. It stores only the differences between two polling intervals and shows them in the Graph
And the Maximum is indeed bigger than the Average. You must have missed the m denoting "milli"
Reinhard

[jon3k]

This is how rrdtool works with COUNTERs. It stores only the differences between two polling intervals and shows them in the Graph
And the Maximum is indeed bigger than the Average. You must have missed the m denoting "milli"
Reinhard

I said that maximum is LESS than the average, not bigger. Didn't miss it, assumed it was short for "million".

But you've answered my question, thanks.