All konnected entities duplicated and renamed

This is the second time I’ve had every konnected entity renamed, breaking every associated automation.

My entity started with the name switch.garage_opener_left, then one day after a few months, it became switch.garage_opener_left_2. This week I think we had a 1 second power interruption (interrupting devices but not HA), and now my entities are renamed again to switch.garage_opener_left_3.

Unfortunately, the Konnected integration no longer has a yaml option. How can I get my old entity names back? How can I keep them from changing?

How does it look like in your router ?, and are you able to set static ip on your device ?

I.E discovery, can cause same “name” device to be added with _2,_3 etc.

So if your router-settings is not “optimal”, and it looses connection for a certain time, discovery will see it as a new (with same name) … as it actually could be :slight_smile:
My temp.sensores after they added the “sensor_identify” it’s all different temp_entities (different sigbee id´s, but same name )
entity_rename

“” This integration uses the SSDP integration, which must be enabled for device discovery to work. If you don’t want to use SSDP you’ll need to manually provide the IP and Port information for each Konnected Panel. The IP/Port info can be found using the Konnected mobile app.
“”

Edit: In regards to wifi/ip devices and UPnP, A device send it’s name and/or ipnumber, but HA can’t read on MAC-address level on external devices( i think) , so it just takes it’s device-name … sadly but true, would be great if HA registered all IP-devices on it’s mac-address, but as HA is a “web” app, it doesn’t dig so deep in the protocol-stack

UPnP uses common [Internet] technologies. It assumes the network must run [Internet Protocol] (IP) and then uses [HTTP], [SOAP] and [XML]on top of IP, in order to provide device/service description.
Responses to search requests are also sent over [UDP], but are instead sent using [unicast] (known as HTTPU). UPnP uses UDP due to its lower overhead in not requiring confirmation of received data and re-transmission of corrupt packets …
… which in short terms means, HA ask “anyone here ?”, devices answer (i am"with it’s ip-number", my name is …)

I have DHCP reservations for all my devices. I see my issue now:

            {
                "area_id": null,
                "capabilities": {},
                "config_entry_id": "0aad06b36894ec2a8980ee2388db001e",
                "device_class": null,
                "device_id": "a4485c027def42d295955d5fbff18c7d",
                "disabled_by": null,
                "entity_category": null,
                "entity_id": "switch.garage_opener_left_2",
                "icon": null,
                "id": "1f10dd6e9060dd43fc69d4087a06a39d",
                "name": null,
                "options": {},
                "original_device_class": null,
                "original_icon": null,
                "original_name": "Garage Opener Left",
                "platform": "konnected",
                "supported_features": 0,
                "unique_id": "ecfabcc0e084-2-200-None-None",
                "unit_of_measurement": null
            },

dup:

            {
                "area_id": null,
                "capabilities": null,
                "config_entry_id": "0aad06b36894ec2a8980ee2388db001e",
                "device_class": null,
                "device_id": "a4485c027def42d295955d5fbff18c7d",
                "disabled_by": null,
                "entity_category": null,
                "entity_id": "switch.garage_opener_left_3",
                "icon": null,
                "id": "a8da6be24fa87986b8b9fd6a10909e3e",
                "name": null,
                "options": {},
                "original_device_class": null,
                "original_icon": null,
                "original_name": "Garage Opener Left",
                "platform": "konnected",
                "supported_features": 0,
                "unique_id": "ecfabcc0e084-2-500-None-None",
                "unit_of_measurement": null
            },

The konnected integration creatively uses the momentary press timing as part of the unique id, so when I adjusted it from 200ms to 500ms it gave it a new ID. I think I can edit my entity registry by hand then and maybe fix the problem. Not sure why the MAC address isn’t enough for the unique id though.

1 Like

Discovery (HA) communication don’t go so deep, it’s at the top of the IP layer, but yes i agree with you, HA should at-least not ADD a Device if it has another with same name, then HA should ask, whether user will “drop” the “new” device, or “rename device/entity”.Lots of small wifi devices have no ways to change names, or running without DHCP

I also saw in the “integration” page for “Konnected” that there were some “optional” choices to add " Name/IP , Port, and setting Discovery to “false” "

The problem is not with the discovery but rather in how they generate the unique id for switches:

        self._unique_id = (
            f"{device_id}-{self._zone_num}-{self._momentary}-"
            f"{self._pause}-{self._repeat}"
        )

The device_id is the mac, and combined with the zone_num it should be unique. No idea why they included the settings in the unique id.

ok, i don’t know what the _momentary is, but it’s seems “wrong” to give id a “new” unique_id because of that, ( And a New entity_id !) when the MAC is a unique as can be.
But i also saw that you can go back to use Yaml , anyway for the issue where router/device goes offline, i think adding the “optional” choices Name/IP + Port number( as it’s unique as-well) i HA config-file, as described under my pic above, and then ADD this in your Router, as-well in Local-DNS ir you have, and if you have an option on the Konnected-device to set static(instead of dhcp-reservation) it would be preferred ( i think), next best is “reservation” as you have( but with both IP, Mac-address,name an port(if possible), works slightly different for different Routers

… i had my big “trial” for my new Router/network-setup, been running since mid December, 24 Wifi devices, 6 wired +14sigbee and my Grid-provider came last friday, to switch “Meter” to modern, they said it would take max20min, i had no power for 3 hours !!!, they had "problems and had to wait for another tech. with replacement parts … my battery-backup for HA-server lasted 2 hours, but all other devices was down for 3 … was i worried !, but everything came up no “dropped” devices

Then if/when you change momentary, pause, wait, i guess you just have to to remove the “old-device” in device-registry

Edit3: PS: i think for safety-reasons, you should stop HA, when edit the core.entity_registry … so HA don’t get the idea to auto-update during your work :slight_smile:

Actually i think the “unique-id” is for “internal” use , an id defined differently depending of device/hardware(what ever they have to go on) :slight_smile: