gistfile1.txt

It's not something I took into consideration because for the most part the emails were being delivered pretty reliably but I can see why it might be a problem so I just had to look into it since I'm one of the contributors for that project and what I found was that the queues are not reliable (RabbitMQ is a little more reliable) as messages can be lost and redis queue (RQ) does not have a way to automatically re-do it, the failed job is flagged and stored for a year and you will have to write code to handle the exception.