My HomeLab Monitoring Rules // feat. Checkmk

Jun 16, 2026 Monitoring

Video write-up

In this video, I clean up the monitoring rules in my homelab with Checkmk. Once you monitor Proxmox hosts, virtual machines, containers, storage, DNS, services, network interfaces, and hardware sensors, your dashboard can quickly turn into a wall of WARN and CRIT states — and then you stop trusting it.

I’ll show the rules and decisions I use to reduce alert noise, ignore services I would never act on, tune thresholds with real data, and keep service discovery clean over time. We’ll also use Checkmk 2.5 features like global search and background activation where they make the workflow faster.

These are the short companion notes for the video. The main idea is simple: monitoring everything is useful, but only if the alerts still mean something.

Labels I Use

I try to target rules with labels instead of individual hostnames.

labels:
  type: proxmox
  platform: baremetal

labels:
  type: server
  platform: vm

Rules from the Video

These are examples from my own homelab setup, as shown in the video. Don’t copy-paste them blindly — use them as a starting point and adjust the labels, services, and thresholds to your own environment.

Ignore k3s/containerd runtime filesystems

Ruleset: Disabled services
Match: Filesystem /run/k3s/containerd/.*
Action: ignore matching services

Ignore transient virtual interfaces

Exclude: docker0, docker_*, br-*, tap*

Virtual server committed memory thresholds

Rule: Virtual server committed memory thresholds
Ruleset: Memory and swap usage on Linux
Target: hosts with label platform:vm
Committed memory warning: 120%
Committed memory critical: 150%

Proxmox Chrony/NTP static check

Rule: Proxmox Chrony/NTP static check
Ruleset: State of NTP time synchronization
Target: hosts with label type:proxmox
Max stratum: 10
Offset warning: 200 ms
Offset critical: 500 ms
Alert delay: 6000 / 7200 seconds

NVMe temperature thresholds

Rule: NVMe temperature thresholds
Ruleset: Temperature
Services: /dev/nvme*, CT2000P510SSD8*, WD_BLACK*
Warning: 60°C
Critical: 85°C

WAN ping thresholds

Rule: WAN ping thresholds
Ruleset: Ping and host check parameters
Target: host tag networking=wan
Round trip warning: 1500 ms
Round trip critical: 3000 ms
Packet loss warning: 80%
Packet loss critical: 100%
Packets: 6
Timeout: 20 seconds