Multicluster Monitoring Solution

[akashb2611]

Hi Everyone,

I have a multiple proxmox cluster deployed in multiple locations. Can you guys please suggest a monitoring solution that can be used to monitor all the clusters on a single dashboard?
If not a single dashboard but a monitoring solution that can be efficient, easy to use, and deploy.
Thanks in advance for your help.

 

[notnulled]

You can try the zabbix proxmox monitoring integration, works very well for us

https://www.zabbix.com/la/integrations/proxmox

 

[UdoB]

to monitor all the clusters on a single dashboard?

While primarily being a Zabbix user - and Zabbix works well! - I found a gem somewhere else: https://grafana.com/grafana/dashboards/19119-proxmox-ve-cluster-flux/

This Dashboard gives me a selection drop-down ("Bucket) to select a cluster. Then I see all all nodes and all VMs in one view. Of course I can build something similar in Zabbix, but that one simply worked without any effort for me to create it manually.

My setup followed a random "InfluxDB and Grafana and Proxmox PVE" instruction found at a random website...

For seeing all clusters in one Dashbboard one would need to create/edit/configure a Dashboard by using those available parts...

Have fun!

 

[LnxBil]

For seeing all clusters in one Dashbboard one would need to create/edit/configure a Dashboard by using those available parts...

... and a gigantic monitor. I' don't know about your environment, yet I cannot see anything in a dashboard with hunderds of VMs in one graph.

 

[UdoB]

... and a gigantic monitor.

Yes! Or more of them

While several 4K-displays are nothing special today the real art is to actually show/see/recognize the important bit.

 

[UdoB]

I cannot see anything in a dashboard with hunderds of VMs in one graph.

Just as an example: the dashboard I mentioned shows one graph with the system load (or CPU usage or ...) of every VM in the cluster. Of course I can not really see which VM is represented by each graphical line. But when one VM goes bonkers I can easily see visually that "something" happens. Then I do mouse-over that line and get a popup telling me the name of the VM and the numerical value of the metric.

On the right side of this kind of graphs is a list of VMs. The elements on this list are clickable and when I click on a specific name it hides all the other VMs. Now I have a clear view

This list is multi-select capable, so I can add other VMs to get comparable information from similar machines by a simple click.


That said... I stay with Zabbix mainly. But Grafana Dashboards - build by people who know how to do that right, not me - are just great.

Have fun!

 

[LnxBil]

But when one VM goes bonkers I can easily see visually that "something" happens.

I'm unsure if I would see this in my background noise. We're running a lot of CI jobs that are heavy, yet they don't run all the time, so we already have spikes that double to quadrupel our normal load just by running those CI jobs. Yet if it fits your needs, great!

But Grafana Dashboards - build by people who know how to do that right, not me - are just great.

Totally agree.

 

[jsterr]

While primarily being a Zabbix user - and Zabbix works well! - I found a gem somewhere else: https://grafana.com/grafana/dashboards/19119-proxmox-ve-cluster-flux/

This Dashboard gives me a selection drop-down ("Bucket) to select a cluster. Then I see all all nodes and all VMs in one view. Of course I can build something similar in Zabbix, but that one simply worked without any effort for me to create it manually.

My setup followed a random "InfluxDB and Grafana and Proxmox PVE" instruction found at a random website...

For seeing all clusters in one Dashbboard one would need to create/edit/configure a Dashboard by using those available parts...

Have fun!

Oh thanks for the link! Last time I checked grafs for Grafana I did not found any for clustered enviroments only for singlehosts! Will try out soon!

 

[Zubin Singh Parihar]

While primarily being a Zabbix user - and Zabbix works well! - I found a gem somewhere else: https://grafana.com/grafana/dashboards/19119-proxmox-ve-cluster-flux/

This Dashboard gives me a selection drop-down ("Bucket) to select a cluster. Then I see all all nodes and all VMs in one view. Of course I can build something similar in Zabbix, but that one simply worked without any effort for me to create it manually.

My setup followed a random "InfluxDB and Grafana and Proxmox PVE" instruction found at a random website...

For seeing all clusters in one Dashbboard one would need to create/edit/configure a Dashboard by using those available parts...

Have fun!

Great one!