Monitoring And Alerts

Healthy homelabs do not stay healthy because their owners are lucky.

They stay healthy because the host is being watched, the important services emit signals, and failures have a way to leave the machine before they turn into a Saturday rebuild.

This subsection keeps those two layers separate on purpose.

• the lightweight manual checks that sharpen your intuition
• the proper Prometheus and Grafana stack for when the lab is mature enough to justify one

Pair this subsection with Email Notifications so alerts have somewhere useful to go.

Start With The Simple Checks

Before building a dashboard, know the host well enough to inspect it directly.

# SSH into Proxmox
ssh root@192.168.50.20
 
# Check system health
proxmox-ve-hello
 
# View system load and memory
top -b -n 1 | head -20
 
# Check disk usage
df -h /
zpool list
zfs list
 
# Monitor GPU
nvidia-smi
 
# Check for errors in logs
tail -50 /var/log/syslog | grep -i error
dmesg | tail -20

That does not replace a monitoring system, but it does teach you what normal looks like before a graph starts telling you stories.

The Stack Shape

The full stack in this section uses:

• node_exporter for host metrics
• nvidia_gpu_exporter for GPU metrics
• a SMART textfile collector for disk health
• Prometheus for scraping and storage
• Grafana for dashboards and alerting
• optional pve-exporter for Proxmox API visibility

In This Subsection

• Prometheus And Grafana Stack On Proxmox - build the monitoring LXC, install exporters, wire Prometheus, and bring Grafana online.
• Dashboards And Alerting On Proxmox - import useful dashboards, validate the targets, and turn raw metrics into alerts that actually matter.

How To Read It

If the host is new or the lab is still small, start with the manual checks above and wait until patterns emerge.

If the host is already running workloads you care about, move straight to Prometheus And Grafana Stack On Proxmox, then finish with Dashboards And Alerting On Proxmox.

Related Topics

• Email Notifications - the SMTP path that makes alerts useful.
• Update And Maintenance - where monitoring data should shape maintenance windows instead of being admired after the fact.
• GPU Passthrough On Proxmox - the host-side groundwork behind the GPU telemetry this subsection watches.
• Secure Service Exposure On Proxmox - the place to decide how Grafana or Prometheus should be exposed, if they should be exposed at all.