Gaps in some graphs

Help
#1

Hi!

I’m having some gaps on overall traffic graphs like that:

image
image896×393 60.9 KB

I only have this issue on routers with high traffic load (more than 60Mbps on a Fast ehternet interface). Those routers are Cisco C3560, Cisco C800 or Juniper SRX210. Anyone could give me some advice on where to look for?

Thanks in advance

#2

Hi,

try this:
https://docs.librenms.org/#Support/Performance/
check IOPS on disk (RRDdata), load on devices and try disable unused modules for devices.

#3

If these devices are using snmp v1 then change to v2c or v3.

#4

I’ve changed from v1 to v2c and I’ve disabled unused modules. The poller time has improved a little from 5s avg to 3s, but I still see the same behaviour. It only happens on the traffic graphs, the other graphs doesn’t have those gaps.

#5

hi,

try disabled other snmp client in device. (For me - that’s work for Cisco SG500). Check load on this device.

#6

Please check here -> https://docs.librenms.org/#Support/FAQ/#why-do-i-see-gaps-in-my-graphs

#7

Thanks Kevin,

I think the correct link is that: https://docs.librenms.org/#Support/FAQ/#faq17

I’ve checket the poller and all the devices are checked within 300 seconds:

image
image1504×106 2.7 KB

I’m also running RRDCached, which improved significatively the IOPS performance.
I only have those gaps on traffic graphs and only on the bussiest interfaces

image

I do not use pdns-recrursor, do you think it worth to use it?

#8

Are you 100% sure you are using v2c now? Those graphs look like it’s v1 to me.

#9

Yes, I’m using v2c but with the default values:

image
image1138×308 10.8 KB

do you recommend to change the Port Association Mode, Max Repeater or Max OIDs?

#10

No those don’t have to be changed.

Maybe check the firmware the device is running, could have a bug.

#11