Homelab Network Monitoring: Docker, Proxmox and NAS in One Dashboard

A homelab grows fast. A NAS, a Proxmox box with a few virtual machines, a Docker host with a dozen containers, a Pi-hole, a switch and some access points. Each has its own web page with its own status, and nobody checks all of them every day. In this video we put all of it on one dashboard, with one alarm list and one place for alerts.

We use DeviceShelf Server. It runs as a container or a package on any always-on machine in the lab. Everything shown is its demo network: a made-up homelab with forty six devices, including a Synology, a QNAP, a Proxmox host and a Docker host. No real addresses.

It starts with the network. Every five minutes the server scans every address, notes which devices answer, on which ports, and how fast. Switches, routers and access points are in the same list as the servers, so you see the whole lab, not just the parts that run an agent.

The group view sorts that list by kind: network gear, storage, servers, computers, media and smart home. That's the quickest way to see whether every piece of infrastructure is up.

Back to the table. Now let's look inside the machines.

The Infrastructure view starts with host health. The server reads CPU, memory, disk, temperature and fan speed from every host that speaks SNMP, which includes Synology, QNAP, TrueNAS and any Linux box with the SNMP service. Each value comes with a twenty four hour curve.

The DiskStation's volume is at ninety one percent, and the server projects it full in about twenty three days. That forecast is the useful part: you learn about a full disk weeks before it happens, not when the backup fails. The storage column shows the NAS's own health report, here healthy.

Switch to seven days to see the week. A CPU curve that spikes every night at the same time is usually a scheduled job, not a problem.

Next, Docker. Give the server read access to the Docker socket, and it lists every container with its image, its state and its health check. Traefik, Vaultwarden and Uptime Kuma are fine.

Two are not. Jellyfin is running, but its health check says unhealthy. And Paperless is in a restart loop: it starts, crashes and starts again. Docker's own restart policy hides that, because the container keeps coming back. The server raises an alarm for both.

Then Proxmox. With a read-only API token, the server reads the cluster, its nodes and every virtual machine and container. The overview shows the node's load and memory, the storage pools, the backup jobs and the latest tasks.

The VMs tab lists every virtual machine. Home Assistant OS and the Docker host are running. The Windows lab machine is stopped, and the task list shows why: its last start failed.

The LXC tab does the same for containers, here Pi-hole and the UniFi controller. Tasks and Backups shows whether last night's backup ran. A homelab without a working backup is a hobby with a deadline.

At the bottom, the syslog receiver. Point your router, switches, access points and Proxmox at the server, and their log lines land here, in one list with the time and the sender. That's where you find the moment the switch port went down.

All of that feeds one alarm list. The full disk, the two containers, the stopped VM and a slow DNS check sit side by side, sorted by how serious they are, each with a thirty day trend. One page to check instead of five.

Setting it up is in Settings, Data sources. Proxmox needs its address and an API token with read rights. Synology, QNAP and VMware have their own sections, and the Home Assistant export is here too. Every connection has a test button.

Under Checks and monitors you set the thresholds: CPU, memory, disk and temperature, how many days ahead the disk forecast should warn, and when certificates and domains are about to expire. Add monitors for the services themselves, with a web page that must answer, or a heartbeat your backup script calls when it's done.

To sum up: one server scans the network and reads host health, Docker, Proxmox and syslog, and everything that goes wrong lands in one alarm list, with alerts to your phone. It runs in Docker on the NAS, on Unraid, on a Pi or on the Proxmox box itself. DeviceShelf Server and the setup guide are linked in the description.