# Ubuntu/Debian
sudo apt update && sudo apt install -y lm-sensors fancontrol

# CentOS/Rocky Linux
sudo dnf install -y lm_sensors

# Verify installation
sensors --version

# Auto-detect all hardware monitoring chips
# This command probes for sensor chips and loads appropriate kernel modules
sudo sensors-detect

# The wizard asks several questions — answer YES to all for a thorough probe
# At the end, it shows which modules to load

# Load the detected modules immediately
sudo service kmod start
# Or load them manually:
sudo modprobe coretemp      # Intel CPU temperatures
sudo modprobe nct6775       # Nuvoton super I/O (common motherboard chip)
sudo modprobe it87          # ITE super I/O chip

# Make modules persistent
sudo sensors-detect --auto   # Non-interactive mode, adds modules to /etc/modules

# Verify modules are loaded
lsmod | grep -E "coretemp|nct|it87|k10temp"

# Display all sensor readings
sensors

# Example output:
# coretemp-isa-0000
# Adapter: ISA adapter
# Package id 0:  +45.0°C  (high = +80.0°C, crit = +100.0°C)
# Core 0:        +42.0°C  (high = +80.0°C, crit = +100.0°C)
# Core 1:        +43.0°C
# Core 2:        +41.0°C
# Core 3:        +44.0°C
#
# nct6775-isa-0290
# Adapter: ISA adapter
# SYSTIN:        +35.0°C
# CPUTIN:        +46.0°C
# fan1:          1200 RPM
# fan2:           800 RPM
# VCORE:          +1.06 V
# +12V:          +12.02 V
# +5V:            +5.02 V

# JSON output for scripting
sensors -j

# Fahrenheit output
sensors -f

# Watch sensors in real time
watch -n 2 sensors

# Show specific chip
sensors coretemp-isa-0000
sensors nct6775-isa-0290

# Create the main configuration file
sudo tee /etc/sensors3.conf > /dev/null <<'EOF'
# System-wide sensor configuration

chip "coretemp-isa-*"
  label temp1 "CPU Package"
  label temp2 "CPU Core 0"
  label temp3 "CPU Core 1"
  label temp4 "CPU Core 2"
  label temp5 "CPU Core 3"

  # Set warning threshold for CPU Package temp
  set temp1_max 80
  set temp1_crit 95

chip "nct6775-isa-*"
  label temp1 "Motherboard"
  label temp2 "CPU (Mobo)"
  label fan1 "CPU Fan"
  label fan2 "Case Fan 1"
  label fan3 "Case Fan 2"
  label in0 "Vcore"
  label in1 "+12V"
  label in2 "+5V"

  # Fan minimum speed alerts (RPM)
  set fan1_min 500
  set fan2_min 300

  # Temperature limits for motherboard sensor
  set temp1_max 55
  set temp1_crit 70
EOF

# Verify the config is valid
sensors -c /etc/sensors3.conf

# Reload with new config
sensors

# Run the fan control configuration wizard
# This must be done on the physical console (not SSH) to avoid disruption
sudo pwmconfig

# The wizard:
# 1. Detects PWM-capable fans
# 2. Tests each fan's controllable range
# 3. Maps fans to temperature sensors
# 4. Generates /etc/fancontrol config

# Generated config example:
# /etc/fancontrol
INTERVAL=10           # Check every 10 seconds
FCTEMPS=hwmon2/pwm1=hwmon2/temp2_input
FCFANS=hwmon2/pwm1=hwmon2/fan1_input
MINTEMP=hwmon2/pwm1=40    # Start increasing speed at 40°C
MAXTEMP=hwmon2/pwm1=70    # Maximum speed at 70°C
MINSTART=hwmon2/pwm1=150  # Minimum PWM to start fan from standstill
MINSTOP=hwmon2/pwm1=100   # PWM at minimum running speed

# Enable and start fancontrol service
sudo systemctl enable fancontrol
sudo systemctl start fancontrol

# Check service status
sudo systemctl status fancontrol

# View fancontrol logs
sudo journalctl -u fancontrol -n 50

# Find PWM control files
ls /sys/class/hwmon/hwmon*/pwm*

# Enable manual control for fan (0=full auto, 1=manual, 2=auto)
echo 1 | sudo tee /sys/class/hwmon/hwmon2/pwm1_enable

# Set fan speed (0=off, 255=full speed)
echo 128 | sudo tee /sys/class/hwmon/hwmon2/pwm1  # ~50% speed

# Restore automatic control
echo 2 | sudo tee /sys/class/hwmon/hwmon2/pwm1_enable

# Script to check temperatures and alert if thresholds exceeded
sudo tee /usr/local/bin/temp-monitor.sh > /dev/null <<'EOF'
#!/bin/bash
ALERT_EMAIL="[email protected]"
CPU_WARN=75   # Warning threshold (°C)
CPU_CRIT=90   # Critical threshold (°C)

# Get CPU package temperature
CPU_TEMP=$(sensors -j 2>/dev/null | python3 -c "
import json, sys
data = json.load(sys.stdin)
for chip, vals in data.items():
    if 'coretemp' in chip or 'k10temp' in chip:
        for sensor, reading in vals.items():
            if 'Package' in sensor or 'Tdie' in sensor:
                for k, v in reading.items():
                    if 'input' in k:
                        print(int(v))
                        exit()
" 2>/dev/null || echo 0)

if [ "$CPU_TEMP" -ge "$CPU_CRIT" ] 2>/dev/null; then
  echo "CRITICAL: CPU temperature ${CPU_TEMP}°C exceeds ${CPU_CRIT}°C threshold!" | \
    mail -s "CRITICAL: CPU Temp Alert on $(hostname)" "$ALERT_EMAIL"
elif [ "$CPU_TEMP" -ge "$CPU_WARN" ] 2>/dev/null; then
  echo "WARNING: CPU temperature ${CPU_TEMP}°C exceeds ${CPU_WARN}°C threshold!" | \
    mail -s "WARNING: CPU Temp Alert on $(hostname)" "$ALERT_EMAIL"
fi
EOF

sudo chmod +x /usr/local/bin/temp-monitor.sh

# Schedule via cron (every 5 minutes)
echo "*/5 * * * * root /usr/local/bin/temp-monitor.sh" | \
  sudo tee /etc/cron.d/temp-monitor

# node_exporter automatically exposes hwmon metrics
# Install node_exporter
wget https://github.com/prometheus/node_exporter/releases/latest/download/node_exporter-*.linux-amd64.tar.gz
tar xzf node_exporter-*.tar.gz
sudo mv node_exporter-*/node_exporter /usr/local/bin/

# Start node_exporter (hwmon collector is enabled by default)
sudo systemctl enable node_exporter
sudo systemctl start node_exporter

# Metrics are available at http://localhost:9100/metrics
# Temperature metrics:
curl -s http://localhost:9100/metrics | grep "node_hwmon_temp"
# node_hwmon_temp_celsius{chip="platform_coretemp_0",sensor="temp1"} 45.0

# Average CPU temperature across all cores
avg(node_hwmon_temp_celsius{sensor=~"temp.*"}) by (instance)

# Fan speeds
node_hwmon_fan_rpm{chip=~".*nct.*"}

# Alert rule: CPU above 80°C for 5 minutes
- alert: HighCPUTemperature
  expr: node_hwmon_temp_celsius{sensor="temp1"} > 80
  for: 5m
  labels:
    severity: warning
  annotations:
    summary: "High CPU temperature on {{ $labels.instance }}"

# Sensors are exposed through sysfs
ls /sys/class/hwmon/

# Read temperature directly (value in millidegrees Celsius)
cat /sys/class/hwmon/hwmon0/temp1_input
# 45000 = 45.0°C

# Convert in bash:
TEMP_MILLI=$(cat /sys/class/hwmon/hwmon0/temp1_input)
echo "Temperature: $((TEMP_MILLI / 1000))°C"

# Check if any modules are loaded
lsmod | grep -E "coretemp|nct|it87|k10temp|acpi"

# Run detect again and check which modules it recommends
sudo sensors-detect --auto

# Manually load common modules
sudo modprobe coretemp   # Intel CPUs
sudo modprobe k10temp    # AMD CPUs (Ryzen, EPYC)
sudo modprobe nct6775    # Common Nuvoton chip

# Compare with IPMI readings if available
ipmitool sdr type Temperature

# Check BIOS-reported temps
sudo dmidecode -t temperature 2>/dev/null

# Cross-reference with turbostat (Intel CPUs)
sudo apt install linux-tools-common && sudo turbostat --show Pkg_J,Busy%,Bzy_MHz,PkgTmp

# Check for errors in the config
sudo fancontrol /etc/fancontrol

# fancontrol requires PWM-capable fans connected to the motherboard headers
# USB fans and fans connected via adapters won't appear
ls /sys/class/hwmon/*/pwm*