We have dozens of MR AP models across our network from 18 to 76. Recently there has been an issue with the MR32s alerting and specifying that the DNS is bad and/or misconfigured. Without changing anything with our internal DNS they still seem to alert (typically several at a time), then resolve on their own. I have tried changing the DNS on them but it is difficult to replicate the circumstances since it happens/resolves randomly. All APs have similar/identical configs but it's only the MR32s that ever have an issue.
I spoke to Meraki support at one point and he stated that the updates on the MR32s needed to be rolled back to be better compatible with the current version of the other APs on the network, but the issue still persists. These APs will eventually be upgraded as most others on our network have, but it cannot happen overnight.
The DNS bad or misconfigured alert can sometimes be a bit of a red herring. I've noticed it can appear when there's drops somewhere between the device and the DNS, just enough that it misses enough DNS replies to raise the alert.
Is there any correlation between the AP's that flag the alert and where they're connected to - common switch or network path?
Apologies for the late reply, but no there is no correlation as far as I can tell. There are various different switches on different sites that have this issue. The only common denominator for the alert is that they only come from our MR32 AP's.