@Clive_A 

Thanks for taking the time to reply

"Are you hosting the controller over the cloud? If yes, you have to worry about the connection stability.

I experience the same thing in the lab as well when the Internet becomes unstable and I have an over-cloud adoption.

But if my controller resides in the LAN, my topology next to my office desk works perfectly for months."

I'm hosting my own software based controller with static IP and full domain name ( Omada.Mydomain.com ).

I have ~7 other sites / routers that never have dropped off my OMADA controller, ONLY the 8411 units drop.

One of the 8411 units is on the very same LAN as the controller, and it drops off the controller.

" The key is when the tunnel or port forwarding adoption is not stable, which will interfere with the transmission of the "keepalive" packets. It would eventually lead to a problem like "Heartbeat Missed" > "Disconnected"."

The Omada controller is at my datacenter sitting behind a 8411 with dual 10G internet connections. I host many services behind that 8411, and none of those have issues. I run a national off-site backup service, and software hates small / random drops. So if there are network issues, that backup software (running on 3 servers behind one of the 8411s) would let me know.  ( I do agree a faulty network can cause this, but if that was the case, I'd see more devices going off-line. I have about 60 or so OMADA devices, it is ONLY the 8411 units that drop, no other OMADA device has dropped)

" I would say from the first day when Omada was introduced, there were reports about this. It was never a fix from us. I remembered that back in 2021 when SDNC V4 was released and the cloud was available for it, it had reported. The guides about the adoption over the Internet were first released during that period, too. "

i've seen this when using a Proxy, like NGNIX. The "default" settings in the proxy don't support constant connection OR constant sending of data. Example, by default, if one tries to upload ~500MB to a server behind Ngnix proxy, it would fail. The Proxy would drop. Need to increase the MAX data and increase WAIT times. My stuff is behind a proxy, but I have configured it to allow constant connections, extended the waits and increased max data rates per proxy cfg. But I doubt its PROXY in my case, as out of ~60+ OMADA devices, ONLY my 2 8411 units keep dropping. Again, could be 1hr, could be 1 month, or something in between.

" You might ask this on the controller page where people discuss the controller. But I would assume that this is mostly a network connectivity issue. "

I was debating where to post this. I thought about doing it on the controller support, but the controller is fine, as all my other devices never drop, it's only the 8411 units that drop.

"If you'd like to discuss this with me, let me know how you adopt and manage them. Diagram, please. "

Yea, that would be cool if I could pick your brain a little on this. I think it might be a setting I made in the 8411's config. I have a few point 2 point VPNs going between my sites. I've noticed once an 8411 drops. I'll use the controller to disable the VPNs, factory reset the 8411, base config it so it can see controller, re join to controller and it'll work. If I leave the VPN's enabled, then rejoin the 8411, it'll work. But at some point after the last reboot to apply configs, it'll drop ( 5 to 30 min). I've enabled one VPN every day till it locks or not. Sometimes, I can get them all up and good for a week or so. Maybe a routing / loop issue?  I'd really hate to delete them from OMADA and start fresh, but that is my next option.

Let me draft a layout. I look forward to your input!

Thanks