Write core state to /run/supervisor & monitor (#2027)

* Write core state to /run/supervisor & monitor

* Add watchdog

* Add log if they getting started

* fix stale comment

* Fix script

* avoid output

* fix tests

* fix lint

* Update rootfs/etc/services.d/watchdog/run

Co-authored-by: Franck Nijhof <git@frenck.dev>

Co-authored-by: Franck Nijhof <git@frenck.dev>

rootfs/etc/services.d/supervisor/run

@@ -1,7 +1,7 @@
 #!/usr/bin/with-contenv bashio
 # ==============================================================================
-# Start Service service
+# Start Supervisor service
 # ==============================================================================
 export LD_PRELOAD="/usr/local/lib/libjemalloc.so.2"
 
-exec python3 -m supervisor
+exec python3 -m supervisor

rootfs/etc/services.d/watchdog/finish

@@ -0,0 +1,8 @@
+#!/usr/bin/execlineb -S1
+# ==============================================================================
+# Take down the S6 supervision tree when Watchdog fails
+# ==============================================================================
+if { s6-test ${1} -ne 0 }
+if { s6-test ${1} -ne 256 }
+
+s6-svscanctl -t /var/run/s6/services

rootfs/etc/services.d/watchdog/run

@@ -0,0 +1,34 @@
+#!/usr/bin/with-contenv bashio
+# ==============================================================================
+# Start Watchdog service
+# ==============================================================================
+declare failed_count=0
+declare supervisor_state
+
+bashio::log.info "Start local supervisor watchdog..."
+
+while [[ failed_count -lt 2 ]];
+do
+    sleep 300
+    supervisor_state="$(cat /run/supervisor)"
+
+    if [[ "${supervisor_state}" = "running"  ]]; then
+
+        # Check API
+        if bashio::supervisor.ping; then
+            failed_count=0
+        else
+            bashio::log.warning "Maybe found an issue on API healthy"
+            ((failed_count++))
+        fi
+
+    elif [[  "close stopping" = *"${supervisor_state}"* ]]; then
+        bashio::log.warning "Maybe found an issue on shutdown"
+        ((failed_count++))
+    else
+        failed_count=0
+    fi
+
+done
+
+basio::exit.nok "Watchdog detect issue with Supervisor - take container down!"