We've had an issue lately that caused an issue. A stack of switches came up with a DNS error on the management of the device, but could still contact the cloud. While the switch could connect to the cloud, it was fozen and not forwarding any packets.
We only stumbled over the issue (a switch that is rarely used) and then sent a reboot command.
It rebooted came up online and started forwarding again. Nothing in logs to indicate an issue at all, simply not responsive.
We'd like to alert on the yellow condition so we can pro-actively reboot the switch.
is there a way to do this, cannot find a way in the doco?
Nope, sorry. There's no way built in that will alert on the "Bad DNS" state. You might be able to do something via the API to poll device state and then generate your own alert, but then you're doing an active poll and not just waiting for an alert.
I think it's a bigger problem that the switch stopped forwarding. It should not do that for "Bad DNS". If you come across that again you should open a ticket with support so they can take a look at what's going on. I've had this happen to lots of switches and never have they stopped forwarding.
I've only had issues like that on 9.x code. Have you upgraded to 10.x? If not, schedule it in as soon as you can
Switch stopped forwarding a week before the last upgrade. Not sure what version it was until then, now on 10.35.