Thats basically the way I do it.

pvesh get /cluster/resources --output-format json-pretty | jq --arg k "lxc/$container_id" -r 'map(select(.id == $k))[].name, map(select(.id == $k))[].mem, map(select(.id == $k))[].maxmem, map(select(.id == $k))[].cpu')

Example using pvesh in proxmox. The data is available, just have to use it. I also prefer barebone approach.

Nice, will take a look into it

Does zabbix use a database continuously polling and storing data or is live data used for indication and/or triggers?

I ran a NSLU2 with custom firmware and a mumble server on it. We used it to talk during online gaming without the need for teamspeak etc.
Played BF3 mostly.

Those where the days

Edit: clarification

Depends, in proxmox using ZFS mirror using two drives it goes degraded when missing one.

But would use it carefully tbh as it is all that’s left over in that case. If there is no external backup.

I just built a Server almost using the same parts. Its the ASRock J5040 board, 16gb ram (yes it works) a 500gb m.2 as system using a PCI adapter , 2x4tb ironwolf as ZFS mirror pool, 350 W power supply all in the node 304 fractal case for 550 euro.

Runs proxmox as hypervisor for VM or Container. As example LXC running container for motioneye, jellyfin, pyload with openvpn and syncthing.

I just built a Server using J5040 board. With 16gb ram (yes it works) a 500gb m.2 as system using a PCI adapter , 2x4tb ironwolf as ZFS mirror pool, 350 W power supply all in the node 304 fractal case for 550 euro. It has 4 sata 6 connectors.

Runs proxmox as hypervisor for VM or Container. As example LXC container for motioneye, jellyfin and syncthing.

I just built a Server using J5040 board. With 16gb ram (yes it works) a 500gb m.2 as system , 2x4tb ironwolf as ZFS mirror pool, 350 W power supply all in the node 304 fractal case for 550 euro.

Runs proxmox as hypervisor for VM or Container. As example LXC container for motioneye and syncthing