Server temperature?

Question

linux

Guys, I have a critical situation, see overheating in the northern room, how can I get information on the state of the server temperature, Server on CENTOS, I have already installed the lm-sensors utility, it displays the following parameters (this is in a slightly cooled server room)
very 2.0s: sensors | grep Core Tue Aug 30 20:19:29 2016
Core 0: +51.0°C (high = +84.0°C, crit = +100.0°C)
Core 1: +46.0°C (high = +84.0°C, crit = +100.0°C)
Core 2: +49.0°C (high = +84.0°C, crit = +100.0°C)
Core 3: +51.0°C (high = +84.0°C, crit = +100.0°C)
Core 0: +50.0°C (high = +84.0°C, crit = +100.0°C)
Core 1: +46.0°C (high = +84.0°C, crit = +100.0°C)
Core 2: +52.0° C (high = +84.0°C, crit = +100.0°C)
Core 3: +52.0°C (high = +84.0°C, crit = +100.0°C)
and in normally cooled
very 2.0s: sensors | grep Core Tue Aug 30 20:20:42 2016
Core 0: +31.0°C (high = +96.0°C, crit = +102.0°C)
Core 1: +30.0°C (high = +96.0°C, crit = +102.0°C)
Core 2: +31.0°C (high = +96.0°C, crit = +102.0°C)
Core 3: +28.0°C (high = +96.0°C, crit = +102.0°C)
Core 4: +37.0°C (high = +96.0°C, crit = +102.0°C)
Core 5: +31.0°C (high = +96.0°C, crit = +102.0°C)
Core 6: +32.0° C (high = +96.0°C, crit = +102.0°C)
Core 7: +30.0°C (high = +96.0°C, crit = +102.0°C
) pull out of the BIOS or something else! Please help

Answer 1

Why are you doing | grep Core?
If you remove this part, then sensors will show the rest of the sensors.

Answer 2

Well, what's wrong with each core? You can at least find the average.
Yes, and the temperature is not yet critical, but when it approaches 80, then yes. So far everything is within the normal range.
Yes, you can withdraw via collectd, for example. Display via influxdb + grafana.

Answer 3

If the server supports IPMI, monitor AMB Temp (ambient temperature) either from an ipmitool-based script or in a monitoring system - https://sergeysl.ru/freebsd-ipmi-monitoring-on-zabbix/

Answer 4

SNMP + Dude of old versions, set, pull out all the necessary parameters, make notifications where necessary. The temperature in the same dude can be calculated by simple manipulations as the average. The output will be a notification, we will eat the necessary statistics, maintain charts, and in general, everything is nice, if you fork out a little and buy something like Equicom PING2 or NetPing IO or the like (the choice is really large) after hanging all this with the necessary sensors, then you can generally watch everything that happens in the server room bringing it all into dude. Let's say the minimum temperature sensor is installed in the room, the server door opening sensor and the humidity sensor can always find out for sure that it's hot in the server room because the air conditioner has died or the cooling of the server's CPU is just dying, someone suddenly went there or some water poured from the top floor.