The best place to capture this is probably on Github as part of a new issue. The “update all” logic is not as robust as it probably needs to be now that people have large quantities of scripts installed. If memory serves the backend only supports updating one script at a time. Shortly after the project management features were released “update all” was added as a client side feature where it just issues update requests for all installed scripts - get enough scripts or enough instances of stalled downloads from GH over WiFi and things start to break.

If after running update all things like the maiden REPL don’t work (which is what connection failures to ws://norns.local:5555/ likely indicate) could be a sign that the system is running out of memory and processes are starting to get killed.