Now most people will obviously know it’s a good idea to monitor the temperature of a Datacenter hall and most people will also want to monitor the temperature of their servers.

If you don’t monitor temperature and the cooling fails, you could come in to work to find your hardware fried and a whole world of pain dealing with the fallout…

There are many ways to do this but if you currently have nothing in place to monitor your server temperature and you don’t have any hardware management packs plugged into vROPs there is still an option.

It’s often easy to forget that not all metrics which vROPs management packs can collect are enabled for collection by default (thanks to Simon Eady for the reminder 🙂 ) and this is the case for the temperature metric in the OOB vSphere adapter.

Here is how you enable it…

  • Navigate to the Administration section
  • Policies
  • Select Policy Library tab
  • Select the policy in which you wish to enable this metric for collection
  • Open the policy for edit.
  • Select the ‘Collect Metrics‘ tab and search for ‘sensors|Temperature|Temp C’
  • Enable the metric under the ‘State‘ column by setting it to ‘Local’ enabled (Tick)
  • It would also be a good idea to enable the ‘Dynamic Threshold’ (if not already enabled) so it can monitor for deviations from expected ranges.
  • Save the policy

EnableTempCMetric

Now we have this enabled:

  • Eenter into the context of a suitable ESXi host
  • Select the ‘All Metrics‘ tab
  • Expand the ‘All Metrics’ option in the metric selector
  • Select ‘Sensor
  • Expand the components and see the Temp C Metrics
  • You can now add them to the Canvas on the right

TemperatureMetric

You can now create some custom symptoms and alerts based on these metrics.

Very quick example…CreateTempweratureSymptom