We have identified the issue coming from our nginx reverse proxy running on Minerva as it seemed to be failing. In the logs we found nginx complaining about the "server names hash bucket size". More on that in this gist. The reboot of our system led to nginx performing a reload after a config change that lead to the error. We increased the bucket size as suggested and got nginx back up again. Our monitoring tools suggest that everything is back online again and we confirmed that. Thanks for your patience and sorry for any inconveniance caused.