Monitoring VMs

[sand-max]

Hi guys!

How to get metrics CPU usage, Memory , Disk IO of VMs from host like on graphs on Proxmox web. I could not find in docs of API

 

[LnxBil]

You could use an external metrics server like InfluxDB+Grafana.

 

[sand-max]

You could use an external metrics server like InfluxDB+Grafana.

Thanks for advice.

I configured influxdb + grafana , but faced with strange issue.
I have 5 nodes in cluster , and I see info in some tables of influxdb from 2 only (pvenode1,4), for example:


In tables cpustat only 2nodes

> select * from cpustat where host='pvenode1' limit 1;
name: cpustat
time avg1 avg15 avg5 cpu cpus ctime host idle iowait nice object sum system used user wait
---- ---- ----- ---- --- ---- ----- ---- ---- ------ ---- ------ --- ------ ---- ---- ----
1591626711000000000 13.63 14.09 13.31 0.476287809895513 24 1591626711.8239 pvenode1 11256880708 77705094 10910 nodes 0 1438166352 5473622640 4035445378 0.00500728110528449
> select * from cpustat where host='pvenode2' limit 1;
> select * from cpustat where host='pvenode3' limit 1;
> select * from cpustat where host='pvenode4' limit 1;
name: cpustat
time avg1 avg15 avg5 cpu cpus ctime host idle iowait nice object sum system used user wait
---- ---- ----- ---- --- ---- ----- ---- ---- ------ ---- ------ --- ------ ---- ---- ----
1591626713000000000 5.17 5.5 5.47 0.103695108669046 40 1591626713.29082 pvenode4 27335224219 11373247 1231 nodes 0 516053101 1486459971 970405639 0.00134891934172128
> select * from cpustat where host='pvenode5' limit 1;

and from tables system it's ok:

> select cpu, nodename from system where nodename='pvenode1' limit 1;
name: system
time cpu nodename
---- --- --------
1591626711000000000 0 pvenode1
> select cpu, nodename from system where nodename='pvenode2' limit 1;
name: system
time cpu nodename
---- --- --------
1591626716000000000 0.0949065728333656 pvenode2
> select cpu, nodename from system where nodename='pvenode3' limit 1;
name: system
time cpu nodename
---- --- --------
1591626715000000000 0.501688951676714 pvenode3
> select cpu, nodename from system where nodename='pvenode4' limit 1;
name: system
time cpu nodename
---- --- --------
1591626713000000000 0.0225705323051401 pvenode4
> select cpu, nodename from system where nodename='pvenode5' limit 1;
name: system
time cpu nodename
---- --- --------
1591626715000000000 0 pvenode5
>

 

[MisterY]

I have one problem with proxmox and grafana as well: diskread and diskwrite doesn't work - it tells me "no data". What's wrong?

 

[willybong]

Does enybody have any ideas about that?

 

[SOLTECSIS - Carles Munyoz]

We use Grafana for metrics and checkmk for monitoring.