Support for an efficient "retry all" failed jobs

[manast]

Introduction

In many situations, for example, let's say that due to a bug in your worker you end with thousands of failed jobs, the bug has been fixed and now you want to retry all these jobs as fast as possible, it is important then to be able to just retry all the jobs that have failed in an efficient way.

Requirements

• An atomic method to move all the failed jobs from the failed set to the wait list.
• Preserve the order so that the jobs end in the wait list in the same order as they were failed.
• The retried jobs must be processed before any existing jobs are already in the wait list.