Yesterday morning I lost connection to the tilt. After a powercycle the tilt data showed again in the UI, with a huge gap of course.
This morning I wasn’t able to open the UI. I also couldn’t ping the raspberry. I had to do a power cylce on the pi the get it running again.
In the error log are a couple of erros. I have no clue what it means. The error log is here: https://termbin.com/scjl
If the pattern continues something wil crash tonight, can’t wait to find out 
I’m indeed seeing powercycles at 17-02 05:05 and 18-02 04:23. Apart from that, the eventbus is intermittently reporting connection timeouts with clients, and the tilt service had some issues at 5:05.
It appears that at least your Spark 3 has a 12V power supply, so it’s not that.
You can graph the wifi signal from your sparks. Does that show any weird spikes?
Beyond this equivalent of a squeaky wheel, there’s not much concrete to go on. Maybe run a brewblox-ctl restart, and keep an eye on it.
I was thinking maybe there was a Tilt-only worldwide EMP event, because my Tilt also went offline yesterday morning.
But in my case it was directly related to upgrading to the latest version of Brewblox. I followed the advice in this post and got it working again.
The config issue was an oversight on our side. We specifically created the docker-compose.shared.yml file to separate managed services from user configuration, and forgot that the Tilt service already made changes to those managed services.
I suspect the connection issue may be caused by a new version of the RabbitMQ docker image. Our release caused people to pull updates and notice it. We’ll have to look into it to be sure.