Shelly devices randomly disconnects

I don’t know what to do with it anymore, but since the May version of HA I have problems with shelly devices, of which I have a lot.
Mainly, all the motion sensors fall out. The device appears disconnected in HA, but in the shelly application on the mobile it is normally visible and working. If I restart HA, the devices reconnect and are visible. After some time, they stop working again in different ways. It’s driving me crazy, because of this, various lights around my house don’t turn on. It all worked fine for years until that May HA release there are issues that persist until now. Does anyone know what to do with it? There is a message in the log:

data: Sleeping device did not update within 3600 seconds interval

It’s a known “issue” with other brands also, of battery-driven wifi Devices, they go into “deep sleep” ( to save battery ), only way you can assure it works ( As in your APP ) is install it in a Cloud-Integration

I’ve been using those shelly devices with HA for years and there was no problem, I didn’t touch anything. The problem started after the May HA update.
Otherwise, I have all shelly motion sensors permanently connected to the USB power supply and of course they have the shelly cloud activated. They even respond to pings, but HA simply stops seeing them out of nowhere…

Maybe you should provide some more info, like which Devices and which Integration, then you might get answers from people who have a “similar” setup

BTW, i doubt it helps on the “deep-sleep” functionality, that you have your Charger connected to the devices all the time (that is not what the Charger is for) , i might even overheat, or toast your battery/device eventually

I’m running HA Core 2024.5.5 and all my Shelly devices are working fine. I am running the latest firmware on all my Shelly devices too.

Are you using the official Shelly integration: Shelly - Home Assistant?

Some things I would check:

  1. Do all the devices have the latest version of the firmware?

  2. Have you enabled debugging to see if there is anything in the debug logs?

  3. I would restore my system back to the version prior to the “May” upgrade to confirm that everything works.

1 Like

You have very few shelly devices, so it is possible that the problem will not show up for you, I have more than 70 of them.
I turned on the debug mode, but I don’t know where exactly I can find the outputs from it?
Yes, of course, I regularly update all device shelly, but I repeat that the devices directly in the shelly application on the mobile are normally available and controllable, it is 100% an HA problem. All you have to do is restart HA and all device shells are immediately online in HA and after a while they gradually drop out, in my opinion there is probably some kind of reconnect or something not working there.

The debug file will automatically download when you turn debug off.

And you can follow this guide for more detailed debugging ideas.

EDIT: Forgot to include the link to the guide… 2024.5+: Tracking down instability issues caused by integrations.

Finally, you can log an issue on GitHub with the log info.

Thank you, but the problem is that the log is 200MB :frowning:

Don’t let it run so long - start when it misbehaves, let it run until you think you have caught enough events, stop debugging.

Well, the problem is solved here: All Shelly Motion 2's become unavailable on a regular basis until the service "Home Assistant Core: Reload Config Entry" is called · Issue #119002 · home-assistant/core · GitHub but somehow no one has solved it yet, interesting.

1 Like

And I forgot to include the link to the debugging guide above (but just updated the post). Bieniu recommended one of the steps in the above guide: enabling HA built-in debug.

If it’s the event loop being blocked, any integration can be doing it. The real world analogy is a noisy neighbor problem. You’ll need them all to be quiet for everything to work as expected.

In that case, it’s likely to have multiple causes, so solving will require fixing every misbehaving integration. While you might get lucky and someone else has the same integration that is causing the problem, and it will eventually be found and fixed, it’s also equally likely that the list is different for many people experiencing the problem, so each case must follow the debug steps so you don’t end up waiting and being disappointed to find it it’s a different integration causing the issue.

1 Like

That’s really cool how great is it built that anything can destroy shelly communication? I thought HA was a robust system :frowning:

That’s why I have robust backups - so I can drop back to a known working state if anything has a negative impact on my system.