You are not setting up an integration you are setting up the nginx addon. Or have I misunderstood your issue?
I had 2 issues:
-
The nginx proxy manager addon would change its Docker IP upon restart and make me edit its whitelisted IP in configuration.yaml - this got solved with the /24 setting suggested above
-
Because the nginx proxy manager addon changes its docker IP, it seems like it can get any IP whatsoever, including the one used by my DeConz addon (I guess it started before the DeConz addon and snatched the IP first). This breaks the link between the DeConz addon and the DeConz integration, since the integration discovers an IP when being set up, so is thus hard-coded to the DeConz addon IP at the time of configuration.
For this problem, inputting the DeConz addon host instead of its IP in the DeConz integration would indeed be great, but I wanna be sure I can do that before deleting the existing DeConz integration 
When going for the manual setup of the DeConz integration, I see it asks for a host and a port, so not 3 items like you mentioned above.
1 Like
Ah, unfortunately, HA and the addons are on separate docker networks (HA is on host, the addons are on a specific “hassio”), so that won’t work, unfortunately.
That trick only works between containers on the same docker network.
EDIT: But as Tom said, there is some magic that would make core-deconz resolve.
1 Like
Shoot, it would have been nice to have that working
Would there be any way, then, to assign a static Docker IP to the DeConz addon? Since the Nginx Proxy Manager addon doesn’t have an associated integration, I wouldn’t mind if it continues to move around to various IPs, but DeConz needs to stay put
See my EDIT, ping core-deconz should work
Yeah, only between addons unfortunately. It’s no good for the integration.
Interesting, ping core-deconz does work, but yeah, I don’t think I can use it for the integration, possibly because I’m missing a place in which to set an extra port, there’s only one host and one port to set.
Mmm… I just have the ssh addon, but that works from HA
$ docker exec -it homeassistant /bin/sh
/config # ping core-ssh
PING core-ssh (172.30.33.0): 56 data bytes
64 bytes from 172.30.33.0: seq=0 ttl=64 time=0.513 ms
64 bytes from 172.30.33.0: seq=1 ttl=64 time=0.421 ms
64 bytes from 172.30.33.0: seq=2 ttl=64 time=0.253 ms
EDIT: Works for core-configurator as well…
Does this help?
1 Like
I think it does!
It looks like I could access that .storage config file to change from the IP to core-deconz there, which I will do if this occurs again.
Reading thru that thread, though, it really does seem like this situation shouldn’t be happening anymore, so maybe it was a one-off (I have restarted the machine various times and never came across this issue, so maybe its mentioned auto-healing just failed for some reason this one time today).
In any case, thank you all
Well, 3 months later, I upgrade HA OS, the machine restarts, the DeConz addon gets a different IP and once again breaks the integration.
Thus, I decided to dedicate a Raspberry Pi to DeConz+Conbee. It has a DHCP-reserved real IP instead of this Docker internal IP nonsense, so I will have hopefully solved the problem for good.
A nice bonus is that with a separate machine for DeConz, I will be able to play with some high-availability setups via another instance for HA 
@johnflorin
Did you ever solved the issue without running deconz on an dedicated PI?
Currently i am migrating all to my odroid with HA and dont want an deticated PI anymore.
Nope, been happily running DeConz on the separate Pi ever since, with 0 issues. Highly recommended!
1 Like
Thanks for your update. It would be nice if that dont happen
@Robban Can you bring light in here why the deconz IP of the integration get’s not updated when there will be assigned another IP to the docker container?
Now i have changed the config, as stated above, to the host core-deconz, and it survieved a few reboots. But my feelings, and @johnflorin experience is, that after an update it get’s messed up again.
thx
If you use default_config which enables discovery protocols the address should automatically update if it matches the conbee serial number to what is registered in the config entry. I know people are having issues with the addon from time to time but I don’t know why. Partially because I don’t have a setup using the addon so a bit harder to debug
I think you got me wrong, not the conbee serial connection get’s lost.
The internal 172 docker ip of deconz changes but the HA integration has the old one.
That was the issue in my case, just after an update/reboot.
So my question is: Will the integration IP get’s updates when the docker ip changes?
Ps.: i can understand that without that setup it’s hard to debug.
thx
Right back at ya! 
The Conbee serial number is used as the unique identifier of the integration. In order to automatically update the IP address of the integration the started container needs to be discovered by home assistant and during discovery the Conbee serial number must be matched to the one registered in the integration.
ok, now got ya
If that discovery matches the conbee identifier, will the host in the file config/.storage/core.config_entries be updated with the correct internal 172… IP, which could be new?
In my case:
{
"entry_id": "xxxxx",
"version": 1,
"domain": "deconz",
"title": "xxxx",
"data": {
**"host": "core-deconz", <-- this one**
"port": 40850,
"api_key": "xxxx"
},
"options": {
"master": true
},
"pref_disable_new_entities": false,
"pref_disable_polling": false,
"source": "user",
"unique_id": null,
"disabled_by": null
},
thx
Yes that is the expected outcome when everything works
1 Like
ok, thanks for the confirmation.
Maybe i had setup the conbee before my update and that was the issue.
Because i had installed deconz before bu migrated my deconz installation before the update.
Now it makes sense to me and i hope the issue does’nt come up again.
thx