[mrtg] Logscale and MaxBytes help

[Max (MRTG list)]

I must be doing something wrong still.

As long as I stay below MaxBytes I have the Y scale of the graph at 280k 
like I want it to, but as soon as I go over MaxBytes (but under AbsMax) 
the Y scale jumps up to the level of actual data.

In the test I just did (made a new graph just for it),
  	Max  	Average  	Current
In 	4476.2 kB/s (1619.0%) 	11.7 kB/s (4.2%) 	4476.2 kB/s (1619.0%)
Out 	87.1 kB/s (31.5%) 	371.0 B/s (0.1%) 	87.1 kB/s (31.5%)

as Soon as my current/max went above 280K, the scale went up to match 
Max (scale went up to 4.8M in this case), rather than the clipping that 
you mentioned and I was hopping for.


FYI:MRTG 2.14.7


section of mrtg.cfg for the test graph...

Title[equinox2]: LAN traffic generate by EQUINOX
PageTop[equinox2]: <H1>EQUINOX traffic</H1>
Target[equinox2]: `/usr/local/bin/runmrtg_equinox.sh eth0`
AbsMax[equinox2]: 125000000
MaxBytes[equinox2]: 276480
Unscaled[equinox2]: dwmy
XSize[equinox2]: 390
Options[equinox2]: unknaszero, withzeroes, logscale
WithPeak[equinox2]: wym



Have I got something wrong in there?
NB:. the reason that MaxBytes is so low compared to AbsMax is that I 
just want to pay attention to when the network is talking to the DSL 
modem, and whenever I copy files etc. over the network to other 
computers this isn't very often and I am happy (and hoping) for this 
just to truncate itself at MaxBytes.



William Owen wrote:
> Try adding AbsMax[interface]: <the actual MaxBytes>.  This will cause
> you to unscale to MaxBytes but not discard things over MaxBytes yet less
> than AbsMax.  This will still influence the percent utilization
> calculations and give you the possibility of having >100% utilization.
> I use this to monitor several 3Mbps cable modems (bandwidth is
> approximate) by setting.
> MaxBytes[interface]: 375000 (3Mbps usable bandwidth)
> AbsMax[interface]: 1250000 (10Mbps max interface bandwidth)
> 
> As I said, I occasionally see 110% utilization and (using Routers2 as a
> front-end) it will clip off the top 10% with unscale set.
> 
> -----Original Message-----
> From: mrtg-bounces at lists.oetiker.ch
> [mailto:mrtg-bounces at lists.oetiker.ch] On Behalf Of Max (MRTG list)
> Sent: Sunday, May 27, 2007 12:46 PM
> To: mrtg at lists.oetiker.ch
> Subject: [mrtg] Logscale and MaxBytes help
> 
> Hi,
> 
> I prefer to use logscale with my graphs, as sometimes there is hardly 
> any traffic, sometimes there is a lot, and the log scale makes it a lot 
> easier for me to tell what is happening.
> 
> The problem is that I want to have a group of mrtg graphs keep the same 
> log scale so that I can compare them all at a glance easily without 
> checking the scales.
> 
> I thought that setting MaxBytes all to the same number and then haveing 
> Unscaled set would work, but if any of the interfaces go over MaxBytes 
> then they report zero, which throws off the graph.
> 
> I then tried MaxBytes, but also put AbsMax with the real maximum of the 
> interface, but now if the data rate goes over MaxBytes but stays within 
> AbsMax, then it reports the high value again.
> 
> its possible I have gotten something wrong, but I think I have them set 
> up right.
> 
> Is there any way to either set the scale when using logscale, or a way 
> of telling mrtg to report any value above MaxBytes as MaxBytes rather 
> than a false/zero reading?
> 
> Hope somebody can help.  Sorry for what is probably a newbie question.