[mrtg] Weired behavior of MRTG

[Alex van den Bogaerdt]

On Mon, Sep 10, 2007 at 06:37:37AM -0700, Peter wrote:
> > Peak hour... as in: a high consumer of bandwidth is present?
> > that's why I asked:
> > Why would such fluctuations be a problem?  Are there devices
> > with gigabit ethernet (or similar) on the network?
> 
> Yes, there are devices with gig interfaces on my network...
> 
> > Many different consumers will create tiny fluctuations which are
> > hardly visible.  Also, they tend to cancel each other out.
> 
> tiny? it's not tiny!! look at the graphs again please! it's about 2.5 - 3gb !!! every day at the same time?  

1: "tiny" are those fluctuations that you do _not_ see.  In my sine wave
   example, those are canceling each other out, leaving nothing more than
   some hardly visible noise in an otherwise flat graph (not zero, but
   relatively smooth at e.g. 3G).

2: The visible fluctuations are not tiny.  But the could be the sum
   of a couple smaller fluctuations which are themselves not quite
   so visible.  In my sine wave example, these are the ones amplifying
   each other resulting in a higher amplitude.

3: Every day at the same time... if there is a problem in MRTG, I
   expect a problem when counters wrap around which will NOT happen
   at the same time every day.

4: If there's a problem in the configuration file, it will most
   likely have to do with the total amount of throughput.  In such
   a case, there will be a significant difference between what MRTG
   shows and what your spreadsheet shows.  Double check your max and
   absmax settings.  And do allow for some slack (upto 10%)

Alex