MRTG graph gaps

[jayjay66]

Hi All,

I'm wondering if anybody can help me please. I've got MRTG up & running and it works well BUT I notice that every once in a while (once every 5-6 hours) there are gaps in the MRTG graphs. What can I do to avoid that? Is it avoidable? What do I do to fix that issue.

Any help much appreciated.

Thanks,
JJ

 

[FaiTHLeSS]

What do you mean by gaps?

Are you sure its not just the scale and a very low amount of traffic going through that device at the time making it look like a gap? ie Graph has hit a high peak then gone down to around 0 making it look like 0 traffic.

Or is the device going off or losing connection to the device you are monitoring?

 

[jayjay66]

I'm sure it's not going to 0 but I don't think that the device is gong off. I don't think that it's losing a connection either. Could it be because the server with MRTG that's monitoring it is "overloaded" ?

please let me know if there's anything i can do

thanks,
jj

 

[lgarner]

All are possibilities. If the monitored device isn't going offline (and nothing between it and the monitoring server is, either), then it might be overloaded and unable to respond.

The monitoring server could also be heavily loaded. Is it charting its own load? That might be a clue.

If the devices are actually being monitored and not simply having statistics gathered, you should be notified if something is amiss.

 

[PalmTest]

@jayjay66

Do you use the "unknaszero" option ?
When the gaps occur, do the graphs have a valid timestamp and are the "Cur In:" and "Cur Out:" values just above zero ?
Do the gaps show in the graph as a gray area ?
What version of MRTG (Perl/Apache) do you use ?