Newbie question - best practices when running on Proxmox nodes

Just testing out Netdata by installing the agent on a single physical node/server running Proxmox VE. It looks like it also reports on most stuff inside LXC containers on that node. I was wondering what best practice is for containers - is it worthwhile installing an agent separately within each container? If so, I’d be grateful to understand why. For example, does installing the agent within individual containers provide any information that wouldn’t be provided through the agent on the physical server? I know there isn’t much visibility into VMs so will install a separate instance of the agent in those, but not sure if it’s needed for containers.

I’m also asking because I’m still trying to figure out how many nodes I’ll need to select the right plan.

More generally, I’d also appreciate it if anyone has some pointers as to best practices for monitoring Proxmox with Netdata - anything else that should be done beyond installing the agent? Any suggestions would be most appreciated.

Hey, @dma0. If you need just CPU, memory, disk, and network usage of containers and VMs - there is no need to install Netdata inside LXC containers/VMs. One Netdata installed on the Proxmox host will do. Just try this setup and see if it will be enough for you.

Just realized - VMs memory usage will be from the host system perspective. It will show allocated as used because that is how the system sees it. Real memory usage can be collected only from within VMs (install Netdata inside).

Thank you. I suppose it would be. On a somewhat related note, would there be an issues later switching from Business to Home Lab?

I do not know this :man_shrugging:

would there be an issues later switching from Business to Home Lab?

@hugo ?

The only point to take into consideration is that the Homelab plan is regulated by a Fair Usage Policy if your use case @dma0 falls into that no issues.

1 Like