What Gets Monitored

Each metric maps to a config key in src/config.py. Toggle them on or off to control which sensors are published to Home Assistant.

Metric	Config key	Default
CPU load	`cpu_load`	enabled
CPU temperature	`cpu_temp`	enabled
Used disk space	`used_space`	enabled
Used disk space on extra paths/drives	`used_space_paths`	disabled
Memory usage	`memory`	enabled
Uptime (timestamp)	`uptime`	enabled
Uptime (seconds)	`uptime_seconds`	disabled
Network data sent/received	`net_io`	enabled
Swap usage	`swap`	disabled
CPU clock speed	`sys_clock_speed`	disabled
CPU voltage	`voltage`	disabled
WiFi signal quality (%)	`wifi_signal`	disabled
WiFi signal strength (dBm)	`wifi_signal_dbm`	disabled
Local IPv4 address	`local_ipv4`	disabled
Local IPv6 address	`local_ipv6`	disabled
External IPv4 address	`external_ipv4`	disabled
External IPv6 address	`external_ipv6`	disabled
RPi 5 fan speed	`rpi5_fan_speed`	disabled
RPi power/throttle status	`rpi_power_status`	disabled
HDD/SSD temperature	`drive_temps`	disabled
SSD SMART status	`ssd_health`	disabled
SSD wear / life used (%)	`ssd_health`	disabled
SSD power-on hours	`ssd_health`	disabled
SSD data written (TB)	`ssd_health`	disabled
APT updates available	`apt_updates`	disabled
Intel GPU render busy (%)	`intel_gpu_render`	disabled
Intel GPU video busy (%)	`intel_gpu_video`	disabled
Intel GPU frequency (MHz)	`intel_gpu_freq`	disabled
Intel GPU power (W)	`intel_gpu_power`	disabled
NVIDIA GPU utilization (%)	`nvidia_gpu_util`	disabled
NVIDIA GPU memory (%)	`nvidia_gpu_mem`	disabled
NVIDIA GPU frequency (MHz)	`nvidia_gpu_freq`	disabled
NVIDIA GPU power (W)	`nvidia_gpu_power`	disabled
NVIDIA GPU temperature (°C)	`nvidia_gpu_temp`	disabled
AMD GPU utilization (%)	`amd_gpu_util`	disabled
AMD GPU memory (%)	`amd_gpu_mem`	disabled
AMD GPU frequency (MHz)	`amd_gpu_freq`	disabled
AMD GPU power (W)	`amd_gpu_power`	disabled
AMD GPU temperature (°C)	`amd_gpu_temp`	disabled
Script update available	`git_update`	enabled
External sensors	`ext_sensors`	disabled

GPU sensors. The installer autodetects your GPU vendor and enables the matching sensors.

Intel uses intel-gpu-tools (intel_gpu_top); it requires root, so values only populate when running as the systemd service.

NVIDIA uses the native pynvml library (nvidia-ml-py) and the NVIDIA driver — no root required.

AMD reads sysfs (/sys/class/drm, in-kernel amdgpu driver) with no extra tools — no root required.

CPU voltage. On a Raspberry Pi this reads vcgencmd measure_volts. On other hosts (e.g. x86 / Ubuntu) it falls back to a Vcore reading from hwmon (lm-sensors), which requires a supported Super-I/O driver — install lm-sensors and run sensors-detect; if sensors then shows a Vcore line it is picked up automatically. Many x86 boxes (especially no-name mini-PCs) have no Super-I/O voltage chip at all, so CPU voltage is simply unavailable there. In that case the sensor reports no value (rather than a misleading 0 V); either disable voltage or set use_availability = True for an explicit "unavailable" state in Home Assistant.

SSD health. The installer autodetects non-rotating drives (NVMe and SATA SSDs) and offers to install smartmontools and enable the ssd_health sensors.

smartctl needs root, so values populate when running as the systemd service, or in cron mode via the passwordless sudoers drop-in the installer adds for smartctl.

NVMe metrics come from the uniform SMART health log; SATA wear and data-written are read from the vendor attribute table and are best-effort (they vary by manufacturer).