where is this setting?
This is because with the native integration you are using the deviceâs APIs via http access.
If you implement login for http (Restrict Access option) on a device previously added to the HA, you are also limiting access to the HA.
Deleting and re-adding the device should fix the problem, with integration requiring user / pwd during the process.
You donât need to disable this (security) feature to make the integration work.
Can I join in, I have the same problem, Unavailable Shelly 1âs
I am running 2 RPI4 8GB, one runs the HA OS and one acts as Access Point Hotspot (it is a offline setup, I have no wired internet).
HA is version2022.11.2 and Shelly 1 are firmware 1.12.1.
The Shellyâs have static IP and coloT peer set to the IP og the HA OS.
A nice example is, I have, 2 Shellyâs sitting next to each other and they have a WiFi RSSI of -42 dBM but now one of them have decided to become unavailable so HA can not see it and if I try logging in directly through the Shelly web interface it say the same!
And now a days when a Shelly decides to be unavailable it stays unavailable for ever!
One interesting hack is if I turn off the WiFi on the Access Point for 10 sec and the turn it on again, I will not loose connection to HA but randomly a new Shelly or two will get unavailable and the one that was unavailable gets available again.
What to do winter is coming :0) and it runs all my heating.
Frank
I also have the issue of having unavailable shelly devices but only with a battery powered TRV.
If I wake it up manually it stays available for some hours and then disappears again.
In my house I have about 30 other powered shellys and all run in their own vlan with static ip adresses and CoIoT configured.
Strangely the shelly dw2 (also battery powered) in the same room also works like a charm.
Hi, same situation here, HA on a RasPi and Shelly devices are in different VLANs created on Unify console.
I never had a problem with HA being in the Mgt LAN (192.168.1.x) and Shelly devices in other vlans.
When I added Unifi wifi access points, all broke down, as the mgt LAN was turned into a âprotected LANâ for interaction between the router and the access points.
I then created a 192.168.2.x vlan for HA and things went more complicated for no obvious reasons. Nevertheless, removing all integrations and redoing everything from start 
solved the issue. All Shelly devices (1, 1PM, pro 4PM and EM) worked very smoothly since then and were always well discovered by the integration.
Lately, I bought 2 Shelly H&T and to avoid issues with battery-powered devices, I bought USB-power addons for each, and declared them âpoweredâ in their setup. So theyâre permanently powered. Nevertheless, they still act as if theyâre running on battery, wake up 10s every x minutes, and the integration never caught them. The trick was to unplug/replug them, access their conf via the web interface (not via the app), set the âCoIot peerâ properly. But this was still not sufficient. I could not find any other option than to delete the Shelly integration completely, then re-install it, then after a few minutes, all devices, including H&T were discovered properly.
This week, after 3 months of good work, I lost suddenly the H&T devices. Again, I checked and tried everything, and could not find any other way than deleting the Shelly integration and reinstall it.
So now I make sure I do not change any name or ID that the integration creates for each device, and thatâs the ID I use in every lovelace card or automation. So a delete/reinstall of the integration is painless when needed.
But this was a lot of effort and rework to end up like this. So I hope my experience can save time for others.