sudden spike in Performance Graph Bandwidth MRTG

[zaji_nms]

Dear Expert

Please find the attached weekly and monthly graph

weekly is good but monthly having sudden spike of 20Mbps while our link capacity is just 2Mbps (which is showing correctly in weekly , not going more than that, upto last three weeks okay).

If you see below there is already NOHC=YES

Target[ahm-malc-1_1082]: 1082:mysnmpcomm@ahm-malc-1:::::2
noHC[ahm-malc-1_1082]: yes
SetEnv[ahm-malc-1_1082]: MRTG_INT_IP="" MRTG_INT_DESCR="1-13-30-0"
MaxBytes[ahm-malc-1_1082]: 12500000
Title[ahm-malc-1_1082]: Traffic Analysis for 1082 --
PageTop[ahm-malc-1_1082]: <h1>Traffic Analysis for 1082 --</h1>

We have set MAX in MRTG graph template but this is not the reason as graph showing perfectly 3 weeks out of 4 weeks.

FYI there is two MRTG working as its too big to complete

2014-12-26 17:45:01 -- Started mrtg with config '/etc/mrtg/mrtg.cfg'
2014-12-26 17:45:01 -- Started mrtg with config '/etc/mrtg/mrtg1.cfg'

now I have changed from 5 to 6 for first MRTG as its bigger file
cat /etc/cron.d/mrtg (output truncated)
*/6 * * * * root LANG=C /usr/bin/mrtg /etc/mrtg/mrtg.cfg
*/5 * * * * root LANG=C /usr/bin/mrtg /etc/mrtg/mrtg1.cfg

Regards

[scottwilkerson]

That spike was 4 weeks ago, does the weekly graph for that particular week show a spike?

Also, if the port speed you are monitoring is only 2 Mbps, why do you have the following?

Code: Select all

MaxBytes[ahm-malc-1_1082]: 12500000

a 2Mbps connection should read

Code: Select all

MaxBytes[ahm-malc-1_1082]: 250000

[zaji_nms]

Dear Expert

Yes,

its show spike even you take graph for one day (for that day)

As I said MaxBytes do not think any reason as randomly its giving spike (if this parameter having issue, all the time it will give spike or wrong error).

Why we have MaxByte more than bandwidth, because if link get upgrade, just we have do paper work , not at Nagios MRTG side.

Otherwise we have to do double work for each and every upgrade/downgrade of the link to reflect in MRTG File.

FYI, this is not the issue, may be load on the server? Long MRTG file?
How I know that MRTG files successfully has been read and processed by Nagios except some Warning. We check /var/log/mrtg but its just saying START , not the END log and something like to say /mrtg1.cfg (processed successfully)

Regards

[sreinhardt]

The mrtg log, generally does not contain very much information, as the actual logs as far as mrtg is concerned, are in the rrd files it generates. There are many different reasons this could have happened. Some possibilities off the top of my head, a change was made for a period of time to your nagios monitoring during that period, also this may be an interface on a stacked switch or addon card making it possible for the oid to change. However unless this was a consistent issue, nearly 4 weeks later, its almost impossible to tell what might have causes such a change in stored data.

[zaji_nms]

Dear Expert

If there is some finding, I will share with you (then you have to keep open this thread) or up to you to close, if problem not getting resolved, I will limit the MaxBytes to 2Mbps.

Regards

[abrist]

We will leave the thread open, just wait to reply until you have more info.