Thank you for this, unfortunately the most of all my monitoring is on windows nodes  Thanks anyway.

As temporary workaround (linux only?) you can use lmsensors wrapped with some script.

Something like this:

sensors -A -u coretemp-isa-0000|grep 'temp1_input'|awk '{print $2;}'

 - this will report temperature of the first core in my workstation.

sensors -A -u coretemp-isa-0001|grep 'temp1_input'|awk '{print $2;}'

 - another core.

sensors -A -u coretemp-isa-0002|grep 'temp1_input'|awk '{print $2;}'

 - 3rd, etc.

Update:
GPU temperature for nVidia cards:

nvclock -T|grep temperature|awk '{print $4;}'|cut -dC -f1