I’m starting out in home automation; I chose Gladys Assistant because it’s open-source, local data storage, built-in integrations rather than plugins (correct me if needed), docker as well as documentation that I find clear and well organized. And user-friendly which suits me because I won’t have much time in the coming months.
I think I’ve found most of the answers to my questions on the forum so thanks everyone.
Notably here:
But there are a few left.
I’ll put them here and if needed I’ll make dedicated threads no problem.
What do you mean by manufacturer gateway/bridge, please?
Does being able to do without it mean you can get rid of the manufacturer’s apps? e.g.: eWelink
How is encryption handled between containers and when communicating with the different modules, please? Is it native or do you have to set it up yourself?
How does feature activation work if I use a docker-compose to deploy everything, please?
Thank you for the work and don’t hesitate if you have any questions. Because I absolutely don’t know how to introduce myself.
Great board! Especially given the price of a Raspberry Pi right now, it almost seems like a better alternative
In home automation, manufacturers often sell a device that will handle creating a Zigbee network (one example among others) turnkey, directly connected to the brand’s cloud on the internet.
Xiaomi, for example, lets you buy the gateway, which you connect to your WiFi, and this then allows you to use their official app to control your Xiaomi devices.
So indeed, without the gateway, there’s no connection to the vendor’s cloud, and thus no official app. Also more privacy, but in some cases you’ll lose certain options that are only available via the official app.
Gladys runs in a single container. Other containers are necessary for certain functions like Zigbee. In that case, exchanges are done via MQTT. I don’t have Gladys on hand but I believe the exchanges are well encrypted.
However, Gladys locally only offers HTTP. Gladys Plus offers secure access from the internet, with HTTPS and end-to-end encryption.
To use docker-compose, you then install Gladys in custom mode; the documentation will guide you so you can convert the docker run command into docker compose.
That’s really great the board! Especially given the price of a Raspberry Pi at the moment, it almost seems like a better alternative
I actually thought the same.
I’ll give you some feedback on it if anyone’s interested.
If I’m going to do experimental stuff I might as well make it useful.
Perfect, I understood that correctly and that’s exactly what I wanted.
Okay thanks.
I can switch http to https. That’s well within my abilities, I think, as is setting up a VPN if I want.
I’ll support Gladys financially as soon as possible, because I think it’s a project that deserves it. But right now I’m still in a work-study program and I just spent quite a bit on hardware.
Running Gladys with docker-compose should be fine.
It’s precisely the containers for other functions like zigbee2mqtt or node-red that make me wonder.
Because, if I understand correctly, it’s their activation in Gladys that triggers their installation, right?
So if I want to integrate them into the compose file how do I do that, please?
By the way, is it really useful?
Basically I use a compose file to have a backup of the installation and because I find it easier to manage a multi-container setup.
Also I’m not against making a tutorial afterwards if it can help other people.
With a bit of technical background, switching Gladys’s local interface to HTTPS isn’t really that complicated!
On the forum I’ve already made a tutorial for using the WireGuard VPN if you’re interested. There are other solutions, don’t hesitate to share your feedback anyway
To wrap up on containers, you can create a docker-compose.yml with Gladys, Watchtower and, if you wish, Node-RED. The latter service is not part of Gladys; it’s more of another project recommended for those who want to go a bit further than what Gladys currently allows, while waiting for the integration of new features / compatibility or to better adapt to a very specific need.
Enjoy exploring, and don’t hesitate if you have any questions or feedback
Yes, I had seen @VonOx.
But it’s out of stock from what I saw and I would like the minimal features I need to be up and running by the end of September. After that I’ll have less time.
That’s also why I haven’t offered my help yet. That and the fact that I don’t code yet. XD
Well, I managed to get everything running. I got my temperature sensors and my smart plug working. I’m so happy. ^^
Next steps: switches paired with motion sensors.
However, I haven’t yet found how to trigger an action depending on the day of the week.
By the way, does a self-learning module exist, for example for heating, please?
In any case, thanks for Gladys — I’m really having fun.
Thanks for the detector and the time condition. I should be able to couple it with times.
Next, I’ll first install the switches.
By the way, during my tests yesterday, when I turned on a fan here, I noticed that deleting the scene does not stop the action programmed inside it.
Is that normal behavior or not?
hmm, this has never happened to me. But what might have happened is that the scene remained running for a few minutes.
Let us know if, after a few hours, it’s still the case, and then we can investigate!
No, I think it’s more a matter of scene timing: for example if you asked Gladys to perform an action that takes time, or even to wait.
In my case I have scenes that must run every hour, and others that must wait 10 minutes before continuing.
If your scene used those functions, then you deleted the scene while it was running and Gladys finished the current scene. That’s a hypothesis that could explain why the scene was still working even though you had deleted it.
For the SD case, where you could be penalized is with a large database (greater than 1GB or 2GB). Some actions then become slower, but that’s fairly rare or localized.
For example when deleting a device’s history.
It doesn’t impact the display of the dashboards, which remains super responsive, or the execution of scenes.
