Here is the setup:
Hyperic 4.5.1
36 agents
90% of agents and hyperic server are virtual machines.
About once a month we get an alert storm for all of the Platform availability alerts at 0%.
Upon investigation, everything is always up and running fine.
I have checked time offset, all < 100ms.
I have verified NTPD running with good configuration.
I have verified that VMware timesync is disabled.
We cannot figure out what is causing this.
Anything else we should look at?
Is it possible the is the rare false positive case that is not related to time offset?
Another crazy theory I have involves VMware vMotion and time sync, but I don't know if it is worth pursuing.
I have read that, even with VMware timesync disabled, after a successful vMotion event, VMware syncs the VM with the time on the new host.
Does that sound remotely plausible?