Guest Heartbeat status occasionally goes red than back to green

[tmilner]

Why is guest heartbeat status "burping"? A VM may go along just fine for hours, then the heartbeat property goes red, then within my next 5 minute poll its gone back to green? I'm using the VI API w/ VC 2.5 and ESX 3.5 systems behind them. The guestHeartBeat status is found in the VirtualMachine object.

http://www.vmware.com/support/developer/vc-sdk/visdk25pubs/ReferenceGuide/vim.VirtualMachine.html

Anyone know why this value is so flaky?

Tom

[admin]

We generally check whether the guest operating system is running properly using the property, guestHeartbeatStatus. It is set to red, when no heartbeat is found i.e. guest operating system stopped responding. The status changes to green when Guest responds normally. It is the behaviour of your Guest OS which is being checked by this metric. So flakyness of this value depends on the health of your Guest.

[tmilner]

I'm seeing these false alarms frequently both internally and at my customer's sites. They are tired all of the false alarms.

It's my understanding that the esx server sets guestHeartBeatStatus based upon communication with VMware tools (running within the Guest O/S). Is this correct? If so, does a red condition mean that VMware tools did not respond to ESX, or that VMware tools did not get a response from the O/S, or both, or something else that I'm missing?

And these are definitely false alarms. The condition lasts no more than a minute. An event is generated indicating a "red" state, then shortly thereafter its back to green and may stay that way for hours. I'm not sure what the Guest O/S is in all of my customer's cases, but its likely from Redmond because that's what we're running internally.

Is there something else that I should be monitoring to verify that the guest O/S is running happily. Do you have any suggestions?

Tom