Is it necessary to monitor server CPU temperature, if it was working just fine for a long time, without any monitoring at all?
Q: Should I bother monitoring my CPU temperature?
A:Every CPU, when working, emits certain amount of heat. Simply speaking, the quicker is the CPU, the more cores it has etc., the more heat it radiates when working.
There’s a known limit of temperature a CPU can sustain without taking damage. This high limit is always specified for a given CPU model. If a CPU is exposed to higher temperatures, a variety of negative consequences may follow – resets, life-span reduction, up to CPU stopping functioning completely. Note that CPU not only heats itself, it also heats the motherboard it’s installed into – the thermal damage may happen to both the motherboard and other peripheral devices as well.
There are hardware pieces, such as CPU coolers, designed to prevent CPUs from overheating. Also, every CPU nowadays has built-in thermal sensors; these sensors can be contacted to obtain the current CPU’s temperature and take whatever action could be required.
This is why it is important to know the current CPU temperature and take measures if this temperature is too high. Perhaps it’s time to replace cooler, or re-arrange hardware pieces etc., but it’s better to know of possible problems before they actually happen.
To control the CPU temperature, one can use SpeedFan free software (for Windows), or lm-sensors (for Linux). In either case it is possible to create data sources (via WMI and/or SNMP) providing required data and monitor them using standard IPHost Network Monitor tools.
If your Windows target machine motherboard manufacturer implements WMI in their drivers, you can get CPU temperature from WMI, as in example below:
Note: the value means temperature in tenths of degrees Kelvin, hence if you need Celsius, you need to substract 2732 from the value and divide the result by 10, i.e. tempC = (value – 2732)/10;