Control the device_id property for maximum flexibility - is it possible?

Configuration
1

I’ve recently switched to Zigbee2MQTT from ZHA. So far so good. However, I’m wondering whether I can make my setup more durable.

I noticed that when a device first joins, an entry like this shows up in my zigbee2mqtt configuration.yaml:

devices:
  '0x001788010915XXXX':
    friendly_name: '0x001788010915XXXX'

If after the device joins I change friendly_name to say Phillips Hue Bulb, I can see in HA that its name was updated to Phillips Hue Bulb; however, references to this device’s entities remian with the old ID:

light.0x001788010915XXXX

I can change these entities manually, but that is quite a labor. I then noticed something else. I unpaired the device, and made sure I have the following in my Z2M’s configuration.yaml:

devices:
  '0x001788010915XXXX':
    friendly_name: 'phillips_hue_bedroom_1'

Now when I re-paired the device, it was given a nice ID, and the same applies to its entities:

light.phillips_hue_bedroom_1
sensor.phillips_hue_bedroom_1_linkquality
etc...

Why do I want this?

Suppose tomorrow my database gets corrupted or [[insert apocalyptic scenario here]]. If everything, including entity IDs, is predefined in YAML files which are backed up - I can seemingly get everything up and running pretty quickly again, including automations (which rely on entity_id and sometimes device_id).

Is this kind of configuration possible?

2

If you use Z2M’s UI to rename it, and enable the option to rename the HA entity_id, you’ll get what you’re looking for.

3

If you have the hardware that is capable, have you considered running home assistant as a virtual machine. You can then backup the entire machine if an apocalyptic event occurs. I store a backup of my virtual machine on an off the network backup drive. Then it does not matter what they are named. You reinstall the VM and life is good.

4

Thanks. But having a backup (which is a good practice regardless) is not the same as having a declarative setup.

5

I will respectfully disagree with you in this case. Backing up the VM is like having a backup from a raid drive. It is as if you cloned a hard drive and can swap. All software and configuration remains intact.

6

Thanks Tinkerer. Do you know if something similar is achievable with other HA integrations? For example giving all Shellys IDs in YAML, etc…?

7

Entirely depends on the integration, and requires them to support YAML configuration. If they do, it’ll be explained in the docs for it.