Hardware sensors (lm_sensors)

Why

check_sensors asks lm_sensors whether any channel is in **ALARM** or **FAULT**. That catches fan stoppages and driver fault lines that FTMON's thermal-zone Celsius check never sees. There is no perfdata — the value is the plugin state itself — so FTMON confirms and notifies without a Trends growth profile.

Install

On Ubuntu 24.04:

sudo apt update
sudo apt install monitoring-plugins-basic lm-sensors
sudo sensors-detect --auto   # once per machine; review if unsure
/usr/lib/nagios/plugins/check_sensors

Executable: /usr/lib/nagios/plugins/check_sensors (regular shell script, not a symlink). Upstream: https://www.monitoring-plugins.org/doc/man/check_sensors.html.

Configure

ftmon recipe install sensors

No numeric thresholds. Optional --ignore-fault can be added to argv if FAULT lines from a known-noisy chip should not raise unknown. Confirm cycles are **two** so brief ALARM chatter does not open incidents immediately.

Why there is no Trends profile

The first Nagios line is status text only (SENSORS OK / alarm / fault) with **no** | perfdata. Growth Trends need a numeric series; use the separate temperature recipe (or Metrics on other checks) for Celsius history.

Test

/usr/lib/nagios/plugins/check_sensors
echo "$?"

Exit states are 0 OK, 1 warning (sensors non-zero), 2 critical (line matching ALARM), and 3 unknown (sensors missing or FAULT without --ignore-fault). Fixtures mirror Monitoring Plugins 2.3.5 message text (including the singular SENSOR CRITICAL wording).

ftmon check
ftmon doctor

Security and permissions

No elevation and no network. The plugin runs sensors from a fixed PATH and reads hwmonsysfs via that tool. Install lm-sensors so the binary exists for the FTMON account.

Upstream and licence

Monitoring Plugins 0, GPL-3.0-or-later. FTMON does not redistribute it.

Verified on 2026-07-14 with Monitoring Plugins **2.3.5** (monitoring-plugins-basic on Ubuntu 24.04, Dell XPS 15): SENSORS OK with fans and coretemp present; trusted_executable_path accepts the script.

Registry example

[check.sensors]
# lm_sensors ALARM/FAULT probe from Monitoring Plugins. No numeric thresholds —
# the plugin grades hardware status text from `sensors`.
argv = ["/usr/lib/nagios/plugins/check_sensors"]
protocol = "nagios"
timeout = "10s"

Monitor definition

schema = 1

[monitor]
name = "sensors"
description = "Hardware sensor ALARM/FAULT status via check_sensors (lm_sensors)"
version = 1
enabled = false
platforms = ["linux"]
interval = "60s"
source = "external"

[source_options]
check = "sensors"
entity = "hw:sensors"

[[rule]]
id = "plugin-warning"
group = "sensors"
when = "plugin_state == 1"
severity = "warning"
confirm_cycles = 2
message = "{plugin_message}"

[[rule]]
id = "plugin-critical"
group = "sensors"
when = "plugin_state == 2"
severity = "critical"
confirm_cycles = 2
message = "{plugin_message}"

[[rule]]
id = "plugin-unknown"
group = "check-health"
when = "plugin_state == 3"
severity = "warning"
confirm_cycles = 2
message = "Sensors check is unknown: {plugin_message}"