Ask HN: Rails Shops, how do you deal with Jobs with zero downtime deploys?

We’re deploying with Kamal and want to avoid killing jobs mid-run during deploys. Currently we pause our queues while deploying but with continuous deployment this leads to jobs being paused too often. I’m just curious how other Rails teams are handing this issue.

Or maybe we have the wrong approach entirely?