Gap in graph

Hello, I need your help!
I start in the network administrator job and it’s the first time I use Librenms.
In Librenms, in all or almost graph have gaps.


I have read a lot of documentation, but I have not got the solution.
Librenms turn in a VM.
I use Firefox Version 61.0.1 (64 bits) for seeing the interface web of Librenms.

Thank for your help

  • Are you using Network storage on your VM like a SAN ? Check atop and see if your system is waiting on disk. RRDCached in-particular will cause high IO and your CPU will be always waiting on writing to disk, causing GAPS in the graphs. In my case, moving it from a VM to a standalone box fixed this.

  • Check the health of your VM, CPU and RAM, and increase resources to it, see if the GAPS improve.

  • For that device, check Graphs -> Poller and see if your SNMP polling is spiking at the same times.

@NicoChoq The first thing is to add (if not added already) the librenms itself to monitor.
That way you will be able to check cpu/ram/diskIO in a friendly way.

Second: https://docs.librenms.org/#Support/Performance/ helped me a lot

How many devices are you monitoring? What are your server(s) specs?

RRDCached will make your disk IO to be lower and not higher as @Chas said :stuck_out_tongue:

@TheGreatDoc My experience of rrdcached here, seems to be related to SAN environments Performance & RRDcached

@Chas Ops, you were saying RRDCached + SAN is where you had issues. My fault.

In my case, with VM+Gluster for storage rrdcached works as expected, reducing disk IO.