Skip to content

Home Assistant — Infrastructure Monitoring Plan

Date: 2026-02-24 Context: Proxmox host (oga), Docker VM (jara), NAS — need visibility into CPU, RAM, disk, temperatures from Home Assistant.

Options Compared

Option What It Monitors RAM Cost New Containers Setup Effort
HA System Monitor Docker VM only (CPU, RAM, disk, network) 0 0 Low (built-in)
Proxmox VE HACS Proxmox host + all VMs (CPU, RAM, disk, uptime, status) 0 0 Low (API token + HACS)
Glances Any host (CPU, RAM, disk, network, temps, per-process) ~70MB each 1 per host Medium
Netdata Deep metrics, auto-dashboards 150-300MB each 1 per host Medium
Prometheus + Grafana Everything, custom dashboards 500MB+ 3+ High
MQTT scripts Custom metrics (temps, SMART) ~0 0 Medium (scripting)

Tier 1 — Immediate (zero new containers)

1a. HA System Monitor (built-in)

  • What: CPU usage, memory usage, disk usage, network throughput, last boot, process counts
  • Scope: Docker VM only (where HA runs)
  • Setup: Settings → Integrations → Add → System Monitor
  • Entities created: sensor.processor_use, sensor.memory_use_percent, sensor.disk_use_percent_*, etc.
  • No install needed — built into HA core

1b. Proxmox VE (HACS integration)

  • What: Proxmox host metrics + per-VM status (CPU, RAM, disk, uptime, running/stopped)
  • Scope: MiniPC (oga) host, OPNsense VM, Docker VM
  • Setup:
  • Create API token in Proxmox web UI (Datacenter → Permissions → API Tokens)
  • Install via HACS → Integrations → Search "Proxmox VE"
  • Configure: Proxmox IP (100.78.12.241), port 8006, API token
  • Entities created: sensor.oga_cpu_used, sensor.oga_memory_used, binary_sensor.docker_running, etc.
  • Zero agents — polls Proxmox API directly from HA

Tier 2 — Short-term (lightweight agent on NAS)

2a. Glances on NAS

The NAS isn't a Proxmox VM, so it needs its own monitoring agent. Glances is lightweight (~70MB RAM) and HA has a built-in integration.

# Add to NAS compose (docker/fixed/nas/)
glances:
  image: nicolargo/glances:latest-full
  container_name: glances
  restart: unless-stopped
  pid: host
  volumes:
    - /var/run/docker.sock:/var/run/docker.sock:ro
  environment:
    - GLANCES_OPT=-w
  ports:
    - "61208:61208"
  • HA integration: Settings → Integrations → Add → Glances → host 100.82.77.97, port 61208
  • Entities: CPU, RAM, disk (per mount), network I/O, temperatures, container stats

2b. Glances on Docker VM (optional)

System Monitor covers most Docker VM metrics, but Glances adds per-container stats and temperatures. Only add if System Monitor isn't enough.

Tier 3 — Optional (specific hardware metrics)

3a. MQTT scripts on Proxmox host

Cron job that publishes hardware-specific data to MQTT:

  • CPU package temperature (sensors command)
  • SMART disk health (smartctl)
  • GPU/iGPU stats (if passthrough configured)
# Example cron script (runs every 60s)
#!/bin/bash
TEMP=$(sensors -u | grep temp1_input | head -1 | awk '{print $2}')
mosquitto_pub -h 100.68.63.168 -t "homelab/proxmox/cpu_temp" -m "$TEMP" \
  -u mqtt_user -P mqtt_pass

HA picks these up via MQTT sensor config. Useful for temperatures that Proxmox API doesn't expose.

Implementation Order

Phase Action Can Do Remotely? Status
Now System Monitor integration (browser) Yes Done
Now Proxmox VE HACS integration (browser + Proxmox web UI) Yes Done
Later Glances on NAS (SSH + compose) Yes Done
Optional MQTT scripts on Proxmox (SSH) Yes Pending
Optional Glances on Docker VM Yes Pending

Dashboard Integration

Once sensors are available, add to the Home Overview Dashboard:

Card Data Source Section
Gauge cards System Monitor (CPU, RAM, disk %) Infrastructure Status
Entity cards Proxmox VE (VM status, uptime) Infrastructure Status
Mini-graph-card Glances (NAS disk, network trends) Infrastructure Status
Conditional card Any sensor above threshold Alerts

Prerequisites

  • [x] HACS installed on Home Assistant
  • [x] System Monitor integration added
  • [x] Proxmox API token created (Proxmox web UI)
  • [x] Proxmox VE HACS integration installed and configured
  • [x] Glances container deployed on NAS
  • [x] Glances HA integration configured

Status (2026-03-10): All tiers (1-2) complete. Tier 3 (MQTT scripts) remains optional. VictoriaMetrics + Grafana (Papa) also deployed separately for historical metrics.