Using Mongoid version 3.1. I had a random secondary db instance failure, and noticed that my webservers became quite slow trying to connect to it but repeated failing. When I removed the dead instance from my mongoid.yml and bounced the webserver process, the issues was resolved. I wonder if there's a way to make Mongoid a bit more resilient here, such as not attempting to connect to a secondary node for a given timeout (say 30 minutes) after a series of failures?