This Friday I updated HA to 2023.3.3 and now most of my ESPhome (2023.2.4) shows up as offline. They do work. For example I have a large display with three MAX7219 displays updating energi flow and the values update and a ESP8266 with an IR-blaster to control the A/C-unit. HA shows them as offline. They come and go - sometimes when I check some units are offline and at another moment other units are offline. No network changes have been made since Friday.
If it’s intermittent then my guess is it’s a coincidence and they are having issues connecting to your wifi network. Possibly a new router has been setup next door that is clashing with yours. I’d get one of your ESP8266 devices that currently isn’t working and bring it closer to your AP and see if it comes up. The other thing you could do is fire up a wifi scanner on your 'phone (eg WiFiman or WiFi Analyzer) and walk around and see what the signal strength is like in the locations where you have issues. Another, less likely, cause would be if you have run out of addresses in your dhcp pool. FWIW I like to track both the uptime and wifi signal strength on my devices just for the heck of it eg:
Thank you for your input. This is in a rural area with 150 yards to nearest neighbour so other WiFis are not around. We run our own IoT 2,4 GHz network across the buildings we control and monitor and all devices are connected using statically assigned DHCP allocations. Beside the 9 ESPhome devices we have 30 Shellys in the IoT network.
Also, all the ESPhome works, the control the stuff and they report the values back to HA. HA doesn’t show the entities as “unavailable” even when they are offline in the add-on. From the ESPhome add-on it’s possible to open Logs for the units that are offline.
As the only change is the HA upgrade I will give the HA server a complete reboot and see if that solves anything.
Also, the add-on has nothing to do with HA integration and can be stopped/deleted after the device has been flashed.
That the device show offline in the add-on is solely a mDNS issue, and probably originates within your router.
Easiest way to get around it is to use fixed IP addresses or specify you local domain name in esphome’s network settings (should correspond to the domain name used in your DNS settings in your router)
Full reboot of the VM (and not just a reboot of HA) on HAOS 9.4 running on ESXi 7.0 solved the issue and all devices are now back online and could be upgraded without any issues. ¯\(ツ)/¯