check-health: monitor CPU load

---- ✂️ ----
🧮️📈️ Health warning: CPU load

The average CPU load on MikroTik is at 76%!
---- ✂️ ----
🧮️📉️ Health recovery: CPU load

The average CPU load on MikroTik decreased to 64%.
---- ✂️ ----

check-health

@@ -10,6 +10,8 @@
 :global GlobalFunctionsReady;
 :while ($GlobalFunctionsReady != true) do={ :delay 500ms; }
 
+:global CheckHealthCPULoad;
+:global CheckHealthCPULoadNotified;
 :global CheckHealthLast;
 :global CheckHealthTemperature;
 :global CheckHealthTemperatureDeviation;
@@ -43,6 +45,20 @@
 
 $ScriptLock $0;
 
+:set CheckHealthCPULoad (($CheckHealthCPULoad * 4 + [ /system/resource/get cpu-load ] * 10) / 5);
+:if ($CheckHealthCPULoad > 750 && $CheckHealthCPULoadNotified != true) do={
+  $SendNotification2 ({ origin=$0; \
+    subject=([ $SymbolForNotification "abacus,chart-increasing" ] . "Health warning: CPU load"); \
+    message=("The average CPU load on " . $Identity . " is at " . ($CheckHealthCPULoad / 10) . "%!") });
+  :set CheckHealthCPULoadNotified true;
+}
+:if ($CheckHealthCPULoad < 650 && $CheckHealthCPULoadNotified = true) do={
+  $SendNotification2 ({ origin=$0; \
+    subject=([ $SymbolForNotification "abacus,chart-decreasing" ] . "Health recovery: CPU load"); \
+    message=("The average CPU load on " . $Identity . " decreased to " . ($CheckHealthCPULoad / 10) . "%.") });
+  :set CheckHealthCPULoadNotified false;
+}
+
 :foreach Voltage in=[ /system/health/find where type="V" ] do={
   :local Name  [ /system/health/get $Voltage name  ];
   :local Value [ /system/health/get $Voltage value ];

doc/check-health.md

@@ -12,32 +12,38 @@ Description
 This script is run from scheduler periodically, sending notification on
 health related events:
 
+* high CPU load
 * voltage jumps up or down more than configured threshold or drops below limit
 * power supply failed or recovered
 * temperature is above or below threshold
 
 Note that bad initial state will not trigger an event.
 
-Only sensors available in hardware can be checked. See what your
+Monitoring CPU load works on all devices. Other than that only sensors
-hardware supports:
+available in hardware can be checked. See what your hardware supports:
 
     /system/health/print;
 
 ### Sample notifications
 
+#### CPU load
+
+![check-health notification cpu load high](check-health.d/notification-01-cpu-load-high.avif)
+![check-health notification cpu load ok](check-health.d/notification-02-cpu-load-ok.avif)
+
 #### Voltage
 
-![check-health notification voltage](check-health.d/notification-01-voltage.avif)
+![check-health notification voltage](check-health.d/notification-03-voltage.avif)
 
 #### Temperature
 
-![check-health notification](check-health.d/notification-02-temperature-high.avif)  
+![check-health notification temperature high](check-health.d/notification-04-temperature-high.avif)  
-![check-health notification](check-health.d/notification-03-temperature-ok.avif)
+![check-health notification temperature ok](check-health.d/notification-05-temperature-ok.avif)
 
 #### PSU state
 
-![check-health notification](check-health.d/notification-04-psu-fail.avif)  
+![check-health notification psu fail](check-health.d/notification-06-psu-fail.avif)  
-![check-health notification](check-health.d/notification-05-psu-ok.avif)
+![check-health notification psu ok](check-health.d/notification-07-psu-ok.avif)
 
 Requirements and installation
 -----------------------------

global-functions

@@ -1075,6 +1075,7 @@
 # return UTF-8 symbol for unicode name
 :set SymbolByUnicodeName do={
   :local Symbols {
+    "abacus"="\F0\9F\A7\AE";
     "alarm-clock"="\E2\8F\B0";
     "calendar"="\F0\9F\93\85";
     "chart-decreasing"="\F0\9F\93\89";