So I have a host with 6 nics. 4 of the nics (vmnic2-5) are teamed.
At the end of the day I just want to be alerted if any vmnic drops, and have it clear automatically if it's restored.
I attempted to setup alarms for uplink redundancy lost, connection lost, and redundancy degraded.
Uplink redundancy alarm will alert when 3 out of 4 vmnics go down as expected.
Connection lost alarm will alert when all 4 vmnics go down as expected.
The issue I have is with the degraded alarm. Initially we just used the stock alarm but found that when you shut down 2 nics and then bring one back online the degraded alarm clears, despite the fact that there is still 1 nic down. Same issue when 3 nics are knocked out, as soon as you bring one nic back online the degraded alarm clears.
I've tried setting up 4 different degraded alarms (one for each vmnic set in conditions) and at first that appeared to work.
You can drop 2 links and get two different degraded alarms, but on the 3rd link you get no degraded alarm, only a uplink redundancy lost. If you then continue and drop the final link you get connection lost alarm. At that point, if you bring back the 1st two vmnics you dropped all alarms will clear despite the fact that there is still 2 vmnics down.
It's as if the degraded isn't really stateful. Does anyone have an idea how to just monitor the NIC's in vCenter and not depend on some other value?
I attempted to make a custom alert with the trigger of esx.problem.net.vmnic.linkstate.down but it never triggers so I'm assuming either it isn't supported or I've not created that correctly. I've tried in 4.1 and 5 and had no joy in either.