There are powerful job queue modules for node.js + redis out there, like kue and bull, to name a few. But powerful API comes at a price, they need complex data structure and redis scripts to achieve features such as delayed job, pause/resume and full text search. And since redis doesn't have traditional transaction, ie. no rollback when one of the command failed, and doesn't trigger error in node-redis driver, things can go south without developers noticing. Plus it's difficult to figure out what really happened due to non-intuitive redis data structure.To us, the proper answer is to design around this problem, instead of adding more features, we want a job queue that's barebone, fully tested, easy to inspect, and doesn't hide errors from developers.