I primarily have two types of automation, complex automations (many triggers + conditions etc) and trivial automations.
The trivial automations follow similar patterns:
Typically only 1 trigger
1 or 2 conditions
1 or 2 actions.
There typically isn’t anything complex in these automation some examples might be:
Turn on desk speakers when presence is detected at my desk.
Switch lights on at sunset (with a condition that there is presence).
…
Where there is commonality I might combine a couple of similar automations together, but for the most part they are unrelated, so I end up with a bunch of trivial automations.
I don’t have a big a home, so I think the total number will likely be limited by that, however I am wondering if people use naming conventions or other techniques to keep their automations organized?
PS I am not worried about the complex automations I haven’t really changed them in months, its the trivial automations I am really interested in.
I use a combination of names, descriptions, labels and (for convenience in the UI) categories.
Names on their own are not enough, I find, because my idea of how the whole system works tends to shift over time (it’s five years old now). A name that made perfect sense a couple of years ago may be really obscure today.
Descriptions with revision histories are useful, particularly as short automations often interact (same trigger, perhaps, or calling the same script) and you can include notes about that.
I introduced a labeling system a while ago when I was playing with AI and it has turned out to be quite helpful in it’s own right. Here’s part of the policy document I worked out for the AI - I still refer to it myself:
Class labels
============
Yellow in the UI
Class labels answer the question “what kind of thing is this?”
Every physical device may have a class label (e.g. power_socket) to show what kind of thing it is.
Every entity representing a sensor may have a class label (e.g. temperature_sensor) to show what kind of thing it is.
Properties labels
=================
Orange in the UI
Properties labels answer the question "what attributes does this device have that might affect its use?"
Every physical device may have a properties label (e.g. battery_powered)
Domain labels
=============
Red in the UI
Domain labels answer the question “what real-world concern does this serve?” (e.g. heating, lighting, security).
Every physical device may have one or more domain labels to show what real-world concerns it serves.
Every entity representing a sensor may have one or more domain labels to show what real-world concerns it serves.
Behaviour labels
================
Blue in the UI
Behaviour labels answer the question "Does it belong in any special behaviour sets?"
Entities may have one or more behaviour labels, for example: off_in_reset, override, flag
Meta labels
===========
Green in the UI
Each entity may have one meta label indicating that it belongs to HA internal logic. For example: llm_tool
Integration labels
==================
Grey in the UI.
Integraton labels answer the question "Where does this device or entity come from?"
Most automations only have domain labels.
It’s a sad story. Once you’re got HA running as you want it, there’s nothing left to do except fiddle with the documentation.
I use a combination of Categories and naming conventions. For instance, I have a bunch of trivial automations that set the color and on/off of the notification LEDs on my Z-Wave and Zigbee dimmers/switches. I put them all in the category of LED Notification, and each one’s name begins with LED Notification.
