Server Monitoring for Home Assistant

Single Docker container that monitors your server and publishes to MQTT with Home Assistant auto-discovery.

For Homeassistant Docker running on Linux only!

This is a personal project and I will not be assisting with issues. I will, however, push any updates or changes that I make. I may later also share the code I use for my dashboard.

Features

• Glances - Host-level CPU, RAM, disk, network via web API
• Container monitoring - Status, health, CPU%, memory%, disk size, restarts
• Error detection - Scans container logs for ERROR/CRITICAL (not warnings)
• System updates - Counts available apt updates, trigger upgrades via MQTT
• Kernel errors - Monitors dmesg for I/O and NVMe errors
• Auto-discovery - Sensors automatically appear in Home Assistant

Setup

1. Clone the repository:

   git clone https://github.com/kattcrazy/server-statistics.git
   cd server-statistics

2. Copy .env.example to .env and configure:

   cp .env.example .env
   nano .env  # Edit with your MQTT settings

3. Build and run:

   docker compose up -d --build

4. In Home Assistant:

   • Glances: Settings → Integrations → Add → Glances (host: your-server-ip, port: 61208)
   • MQTT sensors: Automatically discovered under device "Server Monitor"

Configuration

Variable	Required	Default	Description
MQTT_HOST	Yes	-	MQTT broker IP/hostname
MQTT_PORT	No	1883	MQTT broker port
MQTT_USER	Yes	-	MQTT username
MQTT_PASSWORD	Yes	-	MQTT password
TOPIC_PREFIX	No	server	MQTT topic prefix
LOOP_INTERVAL	No	300	Check interval in seconds
TZ	No	America/Los_Angeles	Timezone
GLANCES_PORT	No	61208	Glances web UI port

MQTT Topics

Per-container (auto-discovered)

• {prefix}/containers/{name}/status - up/down
• {prefix}/containers/{name}/health - healthy/unhealthy/none
• {prefix}/containers/{name}/cpu_percent - CPU %
• {prefix}/containers/{name}/mem_percent - Memory %
• {prefix}/containers/{name}/mem_usage - Memory usage string
• {prefix}/containers/{name}/disk_size - Container disk size
• {prefix}/containers/{name}/restart_count - Restart count
• {prefix}/containers/{name}/error_count - Errors in last 5 min
• {prefix}/containers/{name}/last_error - Most recent error (persists)
• {prefix}/containers/{name}/last_error_level - ERROR/CRITICAL/NONE

Server-wide (auto-discovered)

• {prefix}/updates/count - Available apt updates
• {prefix}/updates/status - idle/running/done/failed
• {prefix}/updates/trigger - Publish "run" to trigger apt upgrade
• {prefix}/system/io_error_count - Kernel I/O errors
• {prefix}/system/last_io_error - Most recent I/O error

Home Assistant Entities

MQTT Sensors (auto-discovered)

After starting, these entities auto-appear under device "Server Monitor":

• sensor.*_status, sensor.*_health, sensor.*_cpu, etc. for each container
• sensor.updates_available, sensor.update_status
• sensor.io_errors, sensor.last_io_error
• button.run_server_updates - Triggers apt upgrade

Glances Integration Entities

Add via: Settings → Integrations → Add → Glances (host: your-server-ip, port: 61208)

Entity	Description
sensor.*_cpu_used	CPU usage %
sensor.*_ram_used	RAM usage %
sensor.*_ram_used_percent	RAM usage in bytes
sensor.*_swap_used	Swap usage %
sensor.*_disk_used_*	Disk usage % per mount
sensor.*_disk_free_*	Disk free space per mount
sensor.*_network_rx_*	Network receive rate (bytes/s)
sensor.*_network_tx_*	Network transmit rate (bytes/s)
sensor.*_load_1min	1-minute load average
sensor.*_load_5min	5-minute load average
sensor.*_load_15min	15-minute load average
sensor.*_uptime	System uptime
sensor.*_cpu_thermal	CPU temperature (if available)

Note: Entity names are prefixed with your Glances integration name (e.g., sensor.server_cpu_used)

Requirements

• Docker with docker compose
• MQTT broker (e.g., Mosquitto)
• Home Assistant with MQTT & Glances integration

Credits

• Glances by Nicolas Hennion (@nicolargo) - The excellent cross-platform system monitoring tool that powers the host-level metrics in this project.