How to view thermal information in Linux

Thermal readings matter when a Linux system slows down under load, runs fans at full speed, or shuts itself off unexpectedly. Viewing hardware sensor data shows whether the processor, board, or firmware-reported zones are approaching warning or critical thresholds before changing cooling, firmware, or workload settings.

The sensors command from lm-sensors reads hardware-monitoring data exposed by kernel drivers under /sys/class/hwmon and related thermal interfaces. It presents package, core, fan, voltage, and board readings in a readable terminal format when the kernel has a driver for the hardware.

Sensor coverage depends on the machine, firmware, kernel driver, and runtime environment. Containers and many virtual machines often report no sensors because host hardware is not exposed, while disk temperature normally comes from drive S.M.A.R.T. data rather than lm-sensors.

1. Install lm-sensors if the sensors command is missing.

   $ sudo apt install lm-sensors

   On Fedora and openSUSE, the package is commonly named lm_sensors. Install it with the native package manager for that distribution.

2. Display the current readable sensor report.

   $ sensors
   coretemp-isa-0000
   Adapter: ISA adapter
   Package id 0:  +48.0°C  (high = +100.0°C, crit = +100.0°C)
   Core 0:        +45.0°C  (high = +100.0°C, crit = +100.0°C)
   Core 1:        +46.0°C  (high = +100.0°C, crit = +100.0°C)
   
   acpitz-acpi-0
   Adapter: ACPI interface
   temp1:        +42.0°C  (crit = +105.0°C)

   Look for processor, board, fan, or ACPI thermal lines that match the hardware being checked. Exact chip names and labels vary by driver and system board.

3. Treat missing sensor output as missing exposure rather than a safe temperature state.

   $ sensors
   No sensors found!
   Make sure you loaded all the kernel drivers you need.
   Try sensors-detect to find out which these are.

   This output is common in containers and many virtual machines. On physical hardware, continue only when hardware probing is acceptable.

4. Probe for sensor drivers on a physical host when sensors has no useful readings.

   $ sudo sensors-detect --auto
   # sensors-detect version 3.6.2
   # Kernel: 6.14.0-24-generic x86_64
   ##### snipped #####
   Driver `coretemp':
     * Chip `Intel digital thermal sensor' (confidence: 9)
   ##### snipped #####
   To load everything that is needed, add this to /etc/modules:
   coretemp

   sensors-detect probes hardware buses and may recommend kernel modules. Run it during a maintenance window on production hosts, and review the reported modules before loading drivers.

5. Load a detected sensor module for the current boot.

   $ sudo modprobe coretemp

   Replace coretemp with the module name reported by sensors-detect. Skip this step when sensors already shows the needed readings.

6. Enable the lm-sensors service when detected modules should load at boot.

   $ sudo systemctl enable --now lm-sensors.service

   Some distributions or hardware paths do not need this unit because the relevant drivers are already built in or loaded another way.
   Related: How to manage a Linux service with systemctl

7. Display raw sensor names for monitoring tools or scripts.

   $ sensors -u
   coretemp-isa-0000
   Adapter: ISA adapter
   Package id 0:
     temp1_input: 48.000
     temp1_max: 100.000
     temp1_crit: 100.000
   Core 0:
     temp2_input: 45.000
     temp2_max: 100.000
     temp2_crit: 100.000

   sensors -u uses raw feature names such as temp1_input, which are easier for scripts and collectors to map than localized display labels.

8. List kernel thermal devices exposed through sysfs.

   $ ls /sys/class/thermal
   cooling_device0  cooling_device1  thermal_zone0  thermal_zone1

   The kernel thermal interface exposes thermal zones and cooling devices even when lm-sensors labels are limited. Availability still depends on the platform.

9. Read the type of the thermal zone being inspected.

   $ cat /sys/class/thermal/thermal_zone0/type
   x86_pkg_temp

   Substitute the thermal zone number that exists on the target system. Names such as x86_pkg_temp, acpitz, or vendor-specific labels identify the driver-reported zone.

10. Read the raw thermal zone temperature.

    $ cat /sys/class/thermal/thermal_zone0/temp
    48000

    Thermal zone temperatures are commonly reported in millidegrees Celsius, so 48000 represents about 48°C.

11. Verify live monitoring by refreshing the readable sensor report while load or fan behavior changes.

    $ watch -n 2 sensors

    Press Ctrl+C to stop watch after confirming that the readings update over several refresh intervals.