Help! Overall Traffic not drawing correctly

Hello

I have a dev and production box - same setup and versioning. The Overall Traffic graphs draw absolutely fine on the Dev box.

On the production box, on only a handful of routers, the Overall Traffic is drawing like this:

How can I troubleshoot this - the graph is wrong, can i force it to re-draw?

Thanks in advance!

It’s something to do with the 1.4TB spikes, they aren’t real, whatever is causing it is compromising the scaling of the graph

on the 6 hour view it draws fine, and the traffic is correct.

Any ideas?

Need to enable rrd tune on those interfaces
https://docs.librenms.org/#Support/FAQ/#faq15

1 Like

Great Thanks I’ll take a look

Hi Kevin

I turned on Globally under Global Settings → External Settings → RRDTool Setup.

But, nothing has happened,

If i click on the device and try manually turn on, For the actual device, Edit Device → Misc

All ports still show RRD-Tune off, any ideas?

follow this how to https://docs.librenms.org/#Extensions/RRDTune/
there are extra steps you didn’t do.

Ahh I see, I thought it would just go and do it!

I’ve ran the script now,
/opt/librenms/scripts# php tune_port.php -h 10.9.12.1
Found hostname 10.9.12.1…
Tuning port Gi6/1…
Tuning port Gi6/2…
Tuning port Gi6/3…
Tuning port Gi6/4…
Tuning port Gi6/5…
Tuning port Gi6/6…
Tuning port Gi6/7…
Tuning port Gi6/8…
Tuning port Gi6/9…
Tuning port Gi6/10…
Tuning port Gi6/11…
Tuning port Gi6/12…
Tuning port Gi6/13…
Tuning port Gi6/14…
Tuning port Gi6/15…
Tuning port Gi6/16…
Tuning port Gi6/17…
Tuning port Gi6/18…
Tuning port Gi6/19…
Tuning port Gi6/20…
Tuning port Gi6/21…
Tuning port Gi6/22…
Tuning port Gi6/23…
Tuning port Gi6/24…
Tuning port Gi7/1…
Tuning port Gi7/2…
Tuning port Gi8/1…
Tuning port Gi8/2…
Tuning port Gi12/1…
etc…

Nothing has changed yet - I guess I need to wait 24 hours for the graph to redraw?

Thanks

I don’t think it’s going to back in and re-draw the graphs for you but idk… I think it will fix the graphs after you made the changes from there out.

Thanks, it’s looking better, not sure why there are Gaps in the graph though :frowning:

that’s usually performance issue with the device or could be LibreNMS.
Run discovery debug on that device see if you getting any timeouts.

Also, run through the performance docs…
https://docs.librenms.org/#Support/Performance/

1 Like

A post was split to a new topic: Performance Issues

Hi again just a quick update, we increased the CPU cores to 16 on this VM and now the gaps slightly less now, we will re-set this up on a standalone server with a better CPU so hopefully that will help. :slight_smile:

Though we have a slight issue with the Overall graphs still, some reason the download and upload drawings are identical and this is certainly not the case.

Our Aggregate graphs in the Port menu at the top are absolutely fine, we’ve added custom descriptions to ports and those aggregates are also fine.

Any ideas?

Thanks!

Maybe, read this, https://docs.librenms.org/#Support/FAQ/#faq15 i know they are not “spikey” but the info in the faq could relate to you.

Also are you 1 min polling or 5 min polling?

Thanks Kevin-

5 minutes polling - RRD tuner is turned on - Just strange the overall graphs seem to be symmetrical.

I’m just thinking, does the Overall Traffic include the port channel interface (Po1) as well as the associated physical interfaces within this port channel?

Overall traffic can be little misleading, Allot my graphs look like that too. But if you go Ports and look at all the Ports you see what each port is reporting back in the data flow. The overall graph pulls all that into one big graph.