Refactor the dispatcher job to use Active Job's `rescue_from` API

[chancancode]

We can refactor this rescue block to use Active Job's rescue_from API.

The main motivation for this (making the code more idiomatic is a nice bonus) is to avoid the automatic retry behavior in Sidekiq. It is not usually useful (and very confusing) to retry the commands. According to this it seems like handling errors manually will stop Sidekiq from retrying automatically, as long as we don't call retry_job.