Gaps in graphs


Component | Version
--------- | -------
LibreNMS  | 22.9.0-3-g06c361c2b
DB Schema | 2022_09_03_091314_update_ports_adsl_table_with_defaults (246)
PHP       | 7.4.3
Python    | 3.8.10
Database  | MariaDB 10.3.34-MariaDB-0ubuntu0.20.04.1
RRDTool   | 1.7.2
SNMP      | 5.8

We are running 1 main LibreNMS server and 2 pollers.

I have some custom devices I’m polling, however there are huge gaps in data. They either stop graphing data for 6, 12 or 24 hours. It’s always dead on 00:00am or 12pm they stop and start at 6am or 18:00.

Here is another

I put my customisations into:


It seems to be all the custom graphs/data for a device will lose it’s graph info for that period, but a similar device on a different IP will be ok for that period but suffer the same another time as if they are waiting or there is a limit set.

So they work, but these terrible gaps they all get. Is it my custom definitions and is LibreNMS waiting to finish on some before it starts on others and why always at set times like 12am or 6am/pm.

Please advise


Look at the poller debug output during that timeframe.

It could be that the device SNMP daemon crashed.

Which debug log in /opt/librenms/logs would that be please?

./poller.php -h <hostname> -d is what I mean when I say run the poller with debug output enabled.

Well this is odd. We have many of the same device custom called ‘GDS’ and using your command is seems some are discovered fine and some are not.

So I run ./poller.php -h -d on a working GDS device were all the custom info comes in and 1 that doesn’t.


As you can see the IP on 10.x.43.24 is discovered correctly and missing the custom sensor stuff, the other is ok.

When I compare the poller.php the one that works use see on the right in green it grabs the sensor info, why though?

I have 2 additional pollers, but the above are polled via the main LibreNMS server

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.