Frequent interruptions with all/most devices

Configuration
1

Been running my installation for about half a year already without any issues. HASS and all 3rd party integrations are running on Docker on my NAS (everything on latest version as of today). I have a larger number of integrations connected, namely Ring alarm via Mosquitto/RingMqtt, Ring doorbell, 5 Hue bridges, 16 Reolink cameras, a hot water controller via LocalTuya, a Goodwe inverter and a handful of other devices. I also have HACS installed which was required for some of these integrations.

For about a month now I can see frequent interruptions in the communication with my devices. They can easily be spotted in the history timeline and look like this:

Hass
Hass1170×157 28.2 KB

Nothing really happened here - the gaps happen because the connection was just interrupted somehow.

At first I didn’t care too much but then I realized this is triggering a lot of my automations for nothing. I have a bunch of checks that do something when the state of an entity changes (e.g. when Ring alarm becomes “Armed” or “Unarmed”) and because these interruptions turn all respective entities to state “Unknown” and then back this always triggers the related automations. So I realized I would need a fix for that.

Now of course these kind of problems are hard to find:

  • As I don’t have any other issues with my network I assume that it’s not related to the router. Also some devices incl. the NAS are connected by cable so Wifi problems do not apply there.
  • I remember having the exact same issues when configuring the Ring alarm via Mosquitto/RingMqtt integration and was able to get rid of them switching the Docker network from Bridged to Custom (in other words: by letting the HASS, Mosquitto and RingMqtt container communicate directly in their own subnet instead of through my normal internal network). This setup is still the same but now the problems remain.
  • The interruptions happen with most devices, but not with all of them (e.g. the Reolink cameras don’t show these gaps in the history timeline). Also not all devices show the gaps at the same time. But the appearances look quite random, at least I didn’t spot any pattern yet.
  • Looked through the HASS logs and found some hints like “We found a custom integration xy which has not been tested by Home Assistant. This component might cause stability problems, be sure to disable it if you experience issues with Home Assistant” so I disabled ALL integrations except for the Ring alarm setup as described above. So I only have the Ring alarm devices connected but still see the interruptions. (Note: Actually I disabled all devices from all other integrations as I didn’t see any option how to disable an integration itself, however, I still see the messages in the log, not sure if I’m missing something?)
  • Before disabling the other devices I saw some more log messages like the one described here which are gone now. The problem persists though.

Further down is the content of my current log (with only Ring alarm enabled). Any hints on how to narrow this down?

2024-08-28 02:18:41.110 WARNING (SyncWorker_0) [homeassistant.loader] We found a custom integration opensprinkler which has not been tested by Home Assistant. This component might cause stability problems, be sure to disable it if you experience issues with Home Assistant 2024-08-28 02:18:41.111 WARNING (SyncWorker_0) [homeassistant.loader] We found a custom integration hacs which has not been tested by Home Assistant. This component might cause stability problems, be sure to disable it if you experience issues with Home Assistant 2024-08-28 02:18:41.112 WARNING (SyncWorker_0) [homeassistant.loader] We found a custom integration sonoff which has not been tested by Home Assistant. This component might cause stability problems, be sure to disable it if you experience issues with Home Assistant 2024-08-28 02:18:41.113 WARNING (SyncWorker_0) [homeassistant.loader] We found a custom integration localtuya which has not been tested by Home Assistant. This component might cause stability problems, be sure to disable it if you experience issues with Home Assistant 2024-08-28 02:18:41.113 WARNING (SyncWorker_0) [homeassistant.loader] We found a custom integration alarmo which has not been tested by Home Assistant. This component might cause stability problems, be sure to disable it if you experience issues with Home Assistant 2024-08-28 02:18:45.567 WARNING (MainThread) [homeassistant.components.homeassistant.triggers.numeric_state] Error initializing 'Goodwe Grid Online' trigger: In 'numeric_state' condition: unknown entity sensor.grid_mode_code 2024-08-28 02:18:45.568 WARNING (MainThread) [homeassistant.components.homeassistant.triggers.numeric_state] Error initializing 'Goodwe Grid Offline' trigger: In 'numeric_state' condition: unknown entity sensor.grid_mode_code 2024-08-28 02:18:45.572 WARNING (MainThread) [homeassistant.components.homeassistant.triggers.numeric_state] Error initializing 'Hot Water Too Cold' trigger: In 'numeric_state' condition: unknown entity sensor.t2 2024-08-28 02:19:15.971 WARNING (MainThread) [homeassistant.helpers.service] Referenced entities item.name are missing or not currently available

The last line appears for all disabled entities, maybe 100x. Besides that this is the full log after running for a couple of hours.

2

This is a standard disclaimer which you get if you use any HACS integration.

All your interruptions are happening at exactly the same time - isn’t this the most significant thing about them? And the entities involved are “unknown”, not simply “unavailable”.

I’d be looking at the Docker/NAS.

3

No, as said the interruptions are not happening at the same time - maybe my first screenshot was somewhat misleading but this was taken after disabling all other devices so the values you see there are all entities from the Ring alarm integration - no surprise the gaps happen at the same time. I now re-enabled the other devices and took another screenshot from last night having entities from most of my integrations mixed and this looks as follows:

Ring
Ring2061×260 104 KB

From top to bottom these are LocalTuya, Ring alarm, Reolink camera, Ring doorbell, Hue light, Goodwe inverter, Sonoff switch and OpenSprinkler. You can see that the gaps are not in synch.

Having a closer look myself it seems the integrations having problems are Ring alarm, Ring doorbell and OpenSprinkler. Now for Ring doorbell and OpenSprinkler I can imagine that these are indeed Wifi issues. They are located a bit far away from the router and I assume the mesh network is moving them around trying if another mesh node can provide better connectivity. These interruptions don’t happen a lot and take only a few seconds - guess we can ignore these.

So the problem actually boils down to the MQTT/Mosquitto/RingMqtt/RingAlarm integration. There is no Wifi involved here, the gaps happen a lot and for a longer time. As said there are 3 docker container involved communicating over a dedicated docker subnet. This setup worked perfectly a few months ago. And I cannot see any issues on the NAS, the load for CPU/Memory/HD/Network all look normal and fine. I will open another issue in the RingMqtt integration forum.

4

Btw It does say “Unavailable” when I hover over the gaps with my mouse:
Ring2