Thanks @Mark actually this turned out to be an issue on your end: a “zombie” instance that was left running after your Oct. 27 outage. It did not show up in our dashboard. I contacted support via email and Kurt eventually diagnosed the problem and shut down the instance. However, we still have a leftover issue: our logs are being flooded with messages like this:
[libcluster:fly6pn] unable to connect to :"production@fdaa:0:309a:a7b:4b:e84f:55b5:2"
…which is again the IP of the “zombie” instance. Again I don’t believe this is anything we can resolve on our end. I’ve gone back and forth with your support and hopefully they’re looking into it now.