Thermal Status Degraded - warning email

[Nat Sacks]

We received a warning e-mail from SIM (v5.2) regarding a thermal status degradation on a server (as a result of an A/C failure).

Is there anyway to make the e-mail display what the recorded temperature was along with the threshold it exceeded?

[David Claypool]

That would be meaningless. The important datum is that it was exceeded.

[Nat Sacks]

Due respect David, that isn't a valid answer to my question. ;)

As it goes, it is not meaningless. the level of the threshold breach might determine a course of corrective action.

[David Claypool]

There are from 3 to 32 sensors for temperature in ProLiant servers, all with individually set thresholds based on model and the component being protected. Depending on the model of server and the sensor tripped, you could have results anywhere from 30C to 105C. The value itself is meaningless, only the fact that it tripped.

[Nat Sacks]

David
I understand that there are many sensors that provide temperature readings. However, knowing the extent of the breach might affect how my team choose to react and, based on a recent experience I therefore deem these alerts to be meaningful.

All I wish to know is if it is possible to change the content of the alert e-mail to include which threshold was breached and to what extent?

I appreciate that you've taken the time to reply but you haven't provided me with an answer to my question.

[David Claypool · Accepted solution]

That implies that you could infer something from that data, and you can't. The event should trigger one action: investigate. The only other thing that possibly could help would be this event in conjunction with another event--such as fan failure, and then you would know the problem is in the system. If you were to receive thermal degraded from multiple systems, that might tell you where to investigate--from multiple systems in a rack, maybe a row airflow problem, insufficient airflow in the rack door, or wraparound air. From systems throughout the computer room--check the environment. Regardless, the event should tell you to investigate.

And no, the contents of the event are not user-modifiable.

[Nat Sacks]

thanks for the answer :)

[Nat Sacks]

solved! :o)