How do you monitor your Proxmox nodes?

[Mor H.]

Hi there,

We had some issues recently with processes failing in our Proxmox cluster and us not catching it until a few hours later.

We want a monitoring solution that will monitor:

• Resource usage (CPU/load, RAM, storage (local and shared), network load)
• Log messages requiring attention
• Health of important services and/or processes

We've been looking at Zabbix for this. Did anyone have experience with Zabbix? What's your impression of it?

How are you monitoring your Proxmox nodes? Any recommendations or things to keep in mind?

All help is appreciated.

Thank you!

 

[oguz]

hi,

We've been looking at Zabbix for this. Did anyone have experience with Zabbix? What's your impression of it?

i've used zabbix, it's pretty cool and you can do a lot of stuff with it and it's easy to use with web GUI & covers quite a lot with zabbix-agent.

other people also use check_mk, prtg, nagios or similar softwares so you can check those too.

 

[adamb]

We are a zabbix shop here. We monitor just about every device under the moon with zabbix.

 

[n1nj4888]

I’ve not used Zabbix myself... My setup is only a small home lab cluster and I use both the inbuilt “External Metric Server” (https://pve.proxmox.com/wiki/External_Metric_Server) to post stats (host, VM, LXC etc) to an InfluxDB database and graph/alert from Grafana but you can also install snmpd on the Proxmox hosts (and even the VMs themselves) and poll them with Telegraf to write to InfluxDB and again visualise with something like Grafana...

The influxdata (InfluxDB and Telegraf) and Grafana communities are also very supportive!

 

[vshaulsk]

I use LibreNMS, which has support specifically for Proxmox. Will show all of individual VM traffic just by adding a node and setting the proxmox flag.
- LibreNMS using SNMP and Check_MK

 

[fireon]

We use Check_MK. You can use OpenSource and Enterprise. That is probably the most powerful but also the simplest monitoring that exists. And it is clusterable. Have fun with it.
https://checkmk.de/

 

[DerDanilo]

Netdata is super simple to setup and does its job very well.

 

[janos]

We using Observium for resource and log monitoring (via SNMP and their agent), and regular Nagios for service monitoring (PVE services, etc).

 

[fireon]

regular Nagios

Sorry, but really stone age. We had Nagios for over 8 years. And with check_mk you safe really a lot of time. But if you love inefficiency then you stay with nagios.

 

[Craig St George]

I use Pandorafms has agents or u can do SNMP etc