Cluster metrics, only reporting CPU and memory

[Gray]

Hello,

I have enabled external metrics via Graphite, and all is working. However the only useful metrics it reports is really only the CPU and memory usage. We have 4 nodes in a cluster and it doesn't even separate out nodes.

I've attached what it reports so far, the only useful metrics being CPU and memory.

Is there any further way to separate out or get more metrics from proxmox like disk, network, or even specific node and VM stats?

 

[dcsapak]

what version do you use? (pveversion -v)

also how does your /etc/pve/status.cfg look like?

i set it just up and i got a seperate metrics folder named "proxmox" where all the metrics are

 

[Gray]

what version do you use? (pveversion -v)

also how does your /etc/pve/status.cfg look like?

i set it just up and i got a seperate metrics folder named "proxmox" where all the metrics are

Running 4.4.

graphite:
    server graphite.(mydomain).net
    port 2003
    path proxmox

The config is correct since it is reporting metrics, but only CPU and memory.

 

[dcsapak]

are you sure you enabled the udp server on graphite and you are looking at the right path?
after manually enabling the udp server (which is not enabled by default) my metrics look like this:


and in proxmox i get all the data
e.g. proxmox.qemu.101.nics.tap101i0.netin
like in the screenshot

 

[Gray]

are you sure you enabled the udp server on graphite and you are looking at the right path?
after manually enabling the udp server (which is not enabled by default) my metrics look like this:
View attachment 4680

and in proxmox i get all the data
e.g. proxmox.qemu.101.nics.tap101i0.netin
like in the screenshot

That did it! Thanks. Is there a way we can edit the wiki page for this to mention that proxmox reports stats over UDP and to have the UDP server enabled?