If I had one, which I had copied several times, now I deleted it and I put a PM suggested by Gladys and it’s good, I am connected 
Perfect, you can retest your Tasmota by using the correct IP, user password, and a single Topic.
You will see all publications pass through MQTT.fx
What do you mean by retest Tasmota?
I subscribed to a temperature sensor topic and I see the values arriving on Mqtt.fx.
But I tried to discover my flashed sonoff devices and I still don’t see them in Tasmota (in Mqtt), because in Http I see them well.
Tasmota (in your Sonoff) needs the broker’s IP, username, and password. It seems to be fine since you can see the messages on MQTT.fx
Can you post a message from Tasmota?
Great, glad I could help.
The Tasmota firmware is fantastic in my opinion, it’s the best integration for me after MQTT in Gladys. You can do everything with it!
I think it was the MP’s word that wasn’t right. I need to modify all the MPs of my devices 
For now, my esp 8266 still doesn’t connect.
I greatly appreciate your help and patience 
Thank you very much
Good evening
It works great, much better than Ewelink
Thank you again, have a good evening
Thanks Pierre-Gilles, I think this integration will help some users better understand how to properly set up the equipment on the first try 
I confirm that the UX improvement is available in Gladys Assistant 4.8:
Hello @pierre-gilles
I installed Portainer to see what was running and better control the containers, and I see that the names given are not very « Gladys friendly ». There are
- gladys
- gladys-z2m-zigbee2mqtt
- gladys-z2m-mqtt
But why not
- gladys-watchtower
- gladys-mosquitto
- gladys-node-red
It may seem trivial, but I find that it would be more coherent and bring a unity of name! In the next release?
By the way, gladys should be gladys-master… I’m just saying, I’m not saying anything 
Already, these three seem quite logical to me:
- gladys
- gladys-z2m-zigbee2mqtt
- gladys-z2m-mqtt
For Mosquitto, I agree that we could have put gladys-mosquitto! However, I’m not sure we should change it now that it’s done, it’s a lot of work to migrate the existing setup and in my opinion it will cause more problems than anything else…
For Watchtower, in my opinion, it doesn’t necessarily have a link with Gladys (Watchtower is a separate project, which updates all the containers in your instance, Gladys or not), so it seems logical to me to just put « watchtower ». It is not even launched by Gladys, it’s either you who installed it, or pre-installed by the raspberry pi os gladys image 
For Node-RED, you launched it, so you must have put the name you wanted… If you want to put gladys-node-red, feel free 
I didn’t realize that since it’s a label name and not the name of the image, it wouldn’t have any impact! Indeed for watchtower, but since it’s pre-installed on the gladys image…
For node-red it’s true, but it’s recommended to control what’s not yet controlled by gladys!
In the end, I changed the names directly in portainer
Do you find this more coherent? I’ll let you know if I encounter any malfunctions later
And I’m thinking about the scenario where there would be a bunch of containers running, now we have them all together!
I advise against doing that for containers managed by Gladys, especially the « mosquitto » container. I checked the code, and we rely on the container name to detect its presence in Gladys. It might work for now, but you expose yourself to bugs that are not easy to debug.
Darn it, blast, bugger, carabistouille, a thousand thunders of Brest…I’ll rollback the name of mosquitto right away! What a shame! 
Thanks to you for checking