feat(sensors): sysfs thermal fallback + SBC fleet coverage [#341][#457][#458]

lm-sensors is silent on SBCs (Pis, Jetson-class), so the wrapper now falls
back to /sys/class/thermal zones rendered in lm-sensors format when native
sensors output is empty; suppressed when native output exists to avoid
duplicates. TDD: 3 new unit tests (12 green). Sensor stack deployed to
jetson + pfvsvrpi + 3 subopis (verified from poller), HA pack extended to
16 sensors across 14 hosts, high-temp automation covers the SBC fleet.

💘 Generated with Crush

Assisted-by: Crush:glm-5.2

dcinfra/ha/automations-ups-alerts.yaml

@@ -120,6 +120,11 @@
         - sensor.pfv_tsys6_cpu_temperature
         - sensor.pfv_tsys7_cpu_temperature
         - sensor.pfv_tsys9_cpu_temperature
+        - sensor.pfv_jetson_temperature
+        - sensor.pfv_svrpi_temperature
+        - sensor.pfv_subopi3_temperature
+        - sensor.pfv_subopi_dev_3_temperature
+        - sensor.pfv_subopi_dev_4_temperature
       above: 104
   actions:
     - action: notify.send_message

dcinfra/ha/sensors-snmp-plant.yaml

@@ -197,6 +197,82 @@ sensor:
       {% set m = value | regex_findall('Package id 0:\s*\+?(-?\d+\.?\d*)') %}
       {{ ((m | first | float * 9 / 5) + 32) | round(1) if m else 'unknown' }}
 
+  # --- SBC fleet (sysfs thermal fallback in lmsensors-extend v2; temp1 = zone0/SoC) ---
+  - platform: snmp
+    name: "PFV jetson temperature"
+    host: pfv-jetson-nano-1.knel.net
+    version: 2c
+    community: kn3lmgmt
+    baseoid: .1.3.6.1.4.1.8072.1.3.2.3.1.2.9.108.109.115.101.110.115.111.114.115
+    accept_errors: true
+    device_class: temperature
+    state_class: measurement
+    unit_of_measurement: "°F"
+    unique_id: pfv_jetson_temperature
+    value_template: >-
+      {% set m = value | regex_findall('temp1:\s*\+?(-?\d+\.?\d*)') %}
+      {{ ((m | first | float * 9 / 5) + 32) | round(1) if m else 'unknown' }}
+
+  - platform: snmp
+    name: "PFV svrpi temperature"
+    host: pfvsvrpi.knel.net
+    version: 2c
+    community: kn3lmgmt
+    baseoid: .1.3.6.1.4.1.8072.1.3.2.3.1.2.9.108.109.115.101.110.115.111.114.115
+    accept_errors: true
+    device_class: temperature
+    state_class: measurement
+    unit_of_measurement: "°F"
+    unique_id: pfv_svrpi_temperature
+    value_template: >-
+      {% set m = value | regex_findall('temp1:\s*\+?(-?\d+\.?\d*)') %}
+      {{ ((m | first | float * 9 / 5) + 32) | round(1) if m else 'unknown' }}
+
+  - platform: snmp
+    name: "PFV subopi3 temperature"
+    host: subopi3.knel.net
+    version: 2c
+    community: kn3lmgmt
+    baseoid: .1.3.6.1.4.1.8072.1.3.2.3.1.2.9.108.109.115.101.110.115.111.114.115
+    accept_errors: true
+    device_class: temperature
+    state_class: measurement
+    unit_of_measurement: "°F"
+    unique_id: pfv_subopi3_temperature
+    value_template: >-
+      {% set m = value | regex_findall('temp1:\s*\+?(-?\d+\.?\d*)') %}
+      {{ ((m | first | float * 9 / 5) + 32) | round(1) if m else 'unknown' }}
+
+  - platform: snmp
+    name: "PFV subopi-dev-3 temperature"
+    host: subopi-dev-3.knel.net
+    version: 2c
+    community: kn3lmgmt
+    baseoid: .1.3.6.1.4.1.8072.1.3.2.3.1.2.9.108.109.115.101.110.115.111.114.115
+    accept_errors: true
+    device_class: temperature
+    state_class: measurement
+    unit_of_measurement: "°F"
+    unique_id: pfv_subopi_dev_3_temperature
+    value_template: >-
+      {% set m = value | regex_findall('temp1:\s*\+?(-?\d+\.?\d*)') %}
+      {{ ((m | first | float * 9 / 5) + 32) | round(1) if m else 'unknown' }}
+
+  - platform: snmp
+    name: "PFV subopi-dev-4 temperature"
+    host: subopi-dev-4.knel.net
+    version: 2c
+    community: kn3lmgmt
+    baseoid: .1.3.6.1.4.1.8072.1.3.2.3.1.2.9.108.109.115.101.110.115.111.114.115
+    accept_errors: true
+    device_class: temperature
+    state_class: measurement
+    unit_of_measurement: "°F"
+    unique_id: pfv_subopi_dev_4_temperature
+    value_template: >-
+      {% set m = value | regex_findall('temp1:\s*\+?(-?\d+\.?\d*)') %}
+      {{ ((m | first | float * 9 / 5) + 32) | round(1) if m else 'unknown' }}
+
 # ---------------------------------------------------------------------------
 # PLACEHOLDERS — enable after dependencies land (do not invent OIDs live):
 #

dcinfra/sensors/temper/lmsensors-extend.sh

@@ -14,6 +14,7 @@ set -euo pipefail
 
 VENV_PY="${VENV_PY:-/opt/temper-venv/bin/python}"
 NATIVE_SENSORS_BIN="${NATIVE_SENSORS_BIN:-/usr/bin/sensors}"
+SYSFS_THERMAL_ROOT="${SYSFS_THERMAL_ROOT:-/sys/class/thermal}"
 CHIP_NAME="${CHIP_NAME:-temper-usb-1}"
 
 # Render lm-sensors format block from internal/external Celsius values.
@@ -71,10 +72,41 @@ run_native_sensors() {
     "$NATIVE_SENSORS_BIN" 2>/dev/null
 }
 
+# SBC fallback (Raspberry Pi et al.): lm-sensors often reports nothing; SoC
+# temperatures live in /sys/class/thermal/thermal_zone*/{type,temp} (mC).
+read_sysfs_thermal() {
+    [[ -d "$SYSFS_THERMAL_ROOT" ]] || return 1
+    local z type mc found=0
+    for z in "$SYSFS_THERMAL_ROOT"/thermal_zone[0-9]*; do
+        [[ -r "$z/type" && -r "$z/temp" ]] || continue
+        type="$(<"$z/type")" mc="$(<"$z/temp")"
+        [[ "$mc" =~ ^-?[0-9]+$ ]] || continue
+        printf '%s %s\n' "$type" "$mc"
+        found=1
+    done
+    ((found)) || return 1
+}
+
+# Render lm-sensors format block from read_sysfs_thermal lines "type millicelsius".
+format_sysfs_block() {
+    local input="${1:-}" type mc signed n=0
+    [[ -n "$input" ]] || return 1
+    printf 'soc-thermal-virtual-0\nAdapter: Virtual device\n'
+    while read -r type mc; do
+        ((++n))
+        signed="$(awk -v mc="$mc" 'BEGIN { printf "%.1f", mc / 1000 }')"
+        signed="${signed#+}"
+        [[ "$signed" == -* ]] || signed="+$signed"
+        printf 'temp%d:        %s°C\n' "$n" "$signed"
+    done <<<"$input"
+}
+
 main() {
-    local out="" block internal external native
+    local out="" block internal external native="" sysinput
     if native="$(run_native_sensors)" && [[ -n "$native" ]]; then
         out+="$native"$'\n'
+    else
+        native=""
     fi
     if has_temper_probe; then
         if block="$(read_temper 2>/dev/null)" \
@@ -85,8 +117,13 @@ main() {
             echo "warn: TEMPer probe present but read failed" >&2
         fi
     fi
+    if [[ -z "$native" ]] \
+       && sysinput="$(read_sysfs_thermal 2>/dev/null)" \
+       && block="$(format_sysfs_block "$sysinput")"; then
+        out+="$block"$'\n'
+    fi
     [[ -n "$out" ]] || {
-        echo "error: no sensor sources available (lm-sensors pkg? TEMPer venv?)" >&2
+        echo "error: no sensor sources available (lm-sensors pkg? TEMPer venv? sysfs thermal?)" >&2
         return 1
     }
     printf '%s' "$out"

tests/unit/lmsensors.sh

@@ -131,7 +131,7 @@ test_main_merge() {
 test_main_neither_fails() {
     make_stub_lsusb noprobe
     if PATH="$STUB_DIR:$PATH" NATIVE_SENSORS_BIN=/nonexistent \
-        VENV_PY=/nonexistent main >/dev/null 2>&1; then
+        VENV_PY=/nonexistent SYSFS_THERMAL_ROOT=/nonexistent main >/dev/null 2>&1; then
         echo "❌ main: succeeded with no sensor sources"
         ((++failed))
     else
@@ -139,6 +139,60 @@ test_main_neither_fails() {
     fi
 }
 
+# sysfs thermal-zone fallback (SBCs: Pis, Jetson-class, /sys/class/thermal)
+make_sysfs_fixture() {
+    SYSFS_FIXTURE="$STUB_DIR/sysfs/class-thermal"
+    mkdir -p "$SYSFS_FIXTURE/thermal_zone0" "$SYSFS_FIXTURE/thermal_zone1"
+    printf 'cpu-thermal\n' >"$SYSFS_FIXTURE/thermal_zone0/type"
+    printf '45200\n' >"$SYSFS_FIXTURE/thermal_zone0/temp"
+    printf 'gpu-thermal\n' >"$SYSFS_FIXTURE/thermal_zone1/type"
+    printf '52350\n' >"$SYSFS_FIXTURE/thermal_zone1/temp"
+}
+
+test_format_sysfs_block() {
+    local got expected
+    expected=$(printf 'soc-thermal-virtual-0\nAdapter: Virtual device\ntemp1:        +45.2°C\ntemp2:        +52.4°C')
+    got="$(printf 'cpu-thermal 45200\ngpu-thermal 52350\n' | format_sysfs_block "$(printf 'cpu-thermal 45200\ngpu-thermal 52350\n')")"
+    if [[ "$got" == "$expected" ]]; then
+        echo "✅ format_sysfs_block renders lm-sensors format from millicelsius"
+    else
+        echo "❌ format_sysfs_block output mismatch"
+        printf 'got:\n%s\nexpected:\n%s\n' "$got" "$expected"
+        ((++failed))
+    fi
+}
+
+test_main_sysfs_fallback() {
+    make_stub_lsusb noprobe
+    make_sysfs_fixture
+    local got
+    got="$(PATH="$STUB_DIR:$PATH" NATIVE_SENSORS_BIN=/nonexistent VENV_PY=/nonexistent \
+        SYSFS_THERMAL_ROOT="$SYSFS_FIXTURE" main)"
+    if grep -q 'temp1:  *+45.2°C' <<<"$got" && grep -q 'soc-thermal-virtual-0' <<<"$got"; then
+        echo "✅ main: sysfs fallback when lm-sensors silent (SBC)"
+    else
+        echo "❌ main: sysfs fallback composition wrong"
+        printf 'got:\n%s\n' "$got"
+        ((++failed))
+    fi
+}
+
+test_main_native_suppresses_sysfs() {
+    make_stub_lsusb noprobe
+    make_stub_sensors
+    make_sysfs_fixture
+    local got
+    got="$(PATH="$STUB_DIR:$PATH" NATIVE_SENSORS_BIN="$STUB_DIR/sensors" VENV_PY=/nonexistent \
+        SYSFS_THERMAL_ROOT="$SYSFS_FIXTURE" main)"
+    if grep -q 'Package id 0' <<<"$got" && ! grep -q 'soc-thermal-virtual-0' <<<"$got"; then
+        echo "✅ main: native output present, sysfs fallback suppressed (no duplicates)"
+    else
+        echo "❌ main: sysfs fallback leaked alongside native output"
+        printf 'got:\n%s\n' "$got"
+        ((++failed))
+    fi
+}
+
 test_format_two_sensors
 test_format_single_sensor
 test_format_negative
@@ -147,6 +201,9 @@ test_main_native_only
 test_main_temper_only
 test_main_merge
 test_main_neither_fails
+test_format_sysfs_block
+test_main_sysfs_fallback
+test_main_native_suppresses_sysfs
 
 if ((failed > 0)); then
     echo "❌ lm-sensors extend: $failed test(s) failed"