@NeilR_M 

The switches are using the latest firmware (4.20.17 Build 20260121 Rel.53429).  Removing or changing back to the old NTP server did not fix the problem.  Rebooting the controller also did not solve the problem.  The odd thing is that all the downstream access points were shown as connected. Traffic through the switches was normal and all clients in the Clients window were also shown as connected.  A screenshot of the Devices window is shown below.

I did get things working again this morning by power cycling the two core switches.  After that, the downstream switches did come back online after a short period.  Later, I again changed the NTP server and everything occurred all over again.  Like before, doing a power down/power up on the two core switches fixed everything.  I have just configured a Synology NAS to be the local time server and after configuring the controller to use the local NTP server, the issue did not happen.

  @NeilR_M 

I was using the Cloudfare time server before trying to change to the Google time server.  The FQDNs for the servers were used in the controller setting.  The IP address of the local NTP server is being used now.

The reason for the original change was that I noticed in the NAS devices that I was still using the Google server.  For some reason, the NAS does not sync with Cloudfare and so I decided to no longer use Cloudfare in the controller.  Hence, I made the change to Google and that's when things went wrong.

  @GRL 

I am not using the gateway as a proxy.  The DNS server for the management VLAN is the gateway address and the DNS servers on the WAN interface are 1.1.1.1 and 1.0.0.1.

I am glad someone else was able to confirm this issue is happening.  For now, I will continue to use a local NTP server that avoids Cloudfare and Google.  The local time server (on a Synology NAS) doesn't like Cloudfare and the controller doesn't like Google.  I am syncing the local time server with pool.ntp.org and everything is working smoothly.

This thread might be related.  The switch disconnects were solved by disabling 802.1X and the Radius server.  I did have 802.1X enabled for testing as I just installed a new outdoor access point but ran into problems with the downstream cameras not connecting.  I did not have time to investigate further and disabled 801.1X but I overlooked disabling the Radius server.  Since disabling the Radius server, I have not tested anything.  Perhaps I'll do that when time permits and the network can go down.