folio-uidemo.aws.indexdata.com is back up. See https://folio-org.atlassian.net/browse/FOLIO-552#icft=FOLIO-552 for additional details.

That's the way its supposed to work, however, with all the manual rebuilds occurring recently, the regular job schedule is off a bit. Ultimately, once the back end modules are a little more stable, we'll have a dedicated demo instance which should be more stable and additional instances that consists of the tip of development.

I’m still wondering why a failed jenkins jobs can break our demo web site? If the update fails, the old instance should still be runnable. Or is this a different story?

HAR!

Looking at the error message, it is pretty clear that this is a permissions naming problem – the user that the Ansible script is using to talk to mod-auth doesn't have the correct permissions to do the task. Permission names are still being worked on, I would expect this to be more or less fixed today.