vFoglight & Infrastructure Agent

You can use the Infrastructure agent alongside with the Vmware Performance agent. The key thing is to not duplicate the data that is already collected from the VM agent.

For the Windows System agent and Unix agent you want to turn off the core collections as follows:

Collect CPU Metrics: False

Collect Disk Metrics: False

Collect Memory Metrics: False

Collect Network Metrics: False

The settings are the same for both Windows and UNIX agents.

Additional details and screenshot posted in the Foglight Community: http://communities.quest.com/message/55514

Regards,

Brian Wheeldon

Thank you Brian and Jeffery.

If I use the Infrastructure agent remotely, which rule then alerts if a host goes down/offline?

Doug

I am trying to locate this specific rule.  I have been able to generate the alarm but, can't seem to find the rule using Rule Management. I would like to clone the rule so I can adjust the subject and message to meet our standards.

I would suggest that you use the NetMonitor agent from the legacy OS cartridge for formalized availability monitoring. This agent has a "quick check" mode that should generate an "unavailable" alarm more quickly than the credential event. The rules for this agent are easier to find, clone and modify than the credential alarm.

Regards,

Brian Wheeldon

Hi Doug,

I just tested this by shutting down a monitored VM in my lab.

The alarm I got was a "Credential Management Event":

Cannot establish connection to 10.4.118.110: The remote machine is visible but has refused the connection [Caused by: There was a problem while connecting to 10.4.118.110:22]

Regards,

Brian Wheeldon

I am trying to turn off teh rule stated above. How do I find it  it is causing fatal alarms to be issued to our NOCC  for machines that don't exist anymore. I am using the OS monitor cartridge to monitor Physical server and also monitor services on virtual machines.  This rule has been causing us headaches because the agents that are alerting on don't exist or thier credentials have moved to a new lockbox and don't exist in the lockbox that is issuing the alerts.

The best way to manage this situation is not to disable the "Device_Unavailable" rule for the NetMonitor agent, but rather to configure that NetMonitor agent(s) so that they monitor only devices that exist.

You should review the "Network Device List" for each NetMonitor agent to ensure that:

- all monitored devices exist

- you wish to be alerted if any of these devices becomes unavailable.

The NetMonitor agent is one of the utility agents in the legacy OS cartridge that does not leverage the credential/lockbox mechanism used by the newer Infrastructure cartridge.

Regards,

Brian Wheeldon

I totally forgot to check the net monitor rule to see if these machines are still in the list. I was trying to chase down the credential issue.