Indeed it has probably something to do with RabbitMQ. This happened to me on 2 different production installation of juniper.
You can check the logs of cms celery queue and confirm the issue (on a native install)
/edx/bin/supervisorctl fg edxapp_worker:cms_default_1
I don’t know exactly how to do that on devstack.
If you get repeating errors like ‘connection reset by host’ or ‘bad credentials’. this means RabbitMQ is not configured properly on your installation and you need correct that.
This blog post explains how to manually reconfigure celery and rabbitmq.
Hope this helps you.