Alarm Status Bounces to "Unavailable" Frequently

[theturtle32]

After upgrading from version 4.6.3 of ring-mqtt to the current version ("latest" tag on docker hub), the alarm status is very frequently showing as "Unavailable" briefly in Home Assistant. I took a screenshot so you can see what I mean.

Is this a known bug? Perhaps related to the upgrade process? What troubleshooting steps would you all recommend?

[tsightler]

There are no known issues of this type and nothing in the code around this has changed for pretty much the entirety of the 4.x releases so it's hard to see how it could be related to the upgrade. I suppose it could be crashing and restarting, but I don't see the restart pattern in that. Logs should provide some clue as to what is happening since all online/offline events are logged.

[tsightler]

Or I guess it could be a bug in MQTT.js library, although I've had no other reports of this behavior and I've not seen it at all, however, it does look quite a bit like this issue:

mqttjs/MQTT.js#1362

[tsightler]

You can enable debug level logging on the mosquitto side and on the ring-mqtt side. On ring-mqtt set DEBUG=mqttjs* to see debug output from the MQTTjs library. You should see the keepalive pings (PINGREQ/PINGRESP) and, if you do the same on the mosquitto server, it might provide some additional insight.

Another thing to check, are the two systems running on the same server? Timesync can be weirdly important in this case. Longshot I know, but I thought I would mention it.

[theturtle32]

Yes, they're running on the same server in different containers.

I've resolved the issue, but I'm still not sure what it was. I simplified/revamped the mosquitto container definition in my docker-compose.yml and it started working reliably, so I went through and put it back to the way it was, one line at a time, to try to isolate which change was the problem, and ended up getting the configuration all the way back to the way it was originally without ever reproducing the timeout disconnection issue again.

So, I'm kind of at a loss. I guess I'll chalk it up to some ephemeral Docker networking issue?

[theturtle32]

In case you're curious, this is the current configuration -- several of these lines are unnecessary, and I'll remove them going forward, but wanted to show it as-is. (The config was originally generated by docker-autocompose)

  mosquitto:
     container_name: mosquitto
     image: eclipse-mosquitto:2.0
     logging:
       driver: json-file
       options:
         max-size: 10m
         max-file: "3"
     tty: true
     stdin_open: true
     mac_address: 02:42:ac:11:00:02
     ports:
       - 1883:1883/tcp
       - 9001:9001/tcp
     volumes:
       - /home/ubuntu/mosquitto/data:/mosquitto/data
       - /home/ubuntu/mosquitto/log:/mosquitto/log
       - /home/ubuntu/mosquitto/config:/mosquitto/config:ro
     restart: unless-stopped

[tsightler]

Thanks for the update, glad you got it working!