[Updater] Drop nats and re-do in BullMQ

[ShootingStar91]

Why?

• The nats seems to drop connection between worker/scheduler sometimes and not regain it
• This, or code bugs / debug attempts, can lead to situations where you have millions of student updates queued. These cannot currently be emptied in any other way than probably removing and returning the persistent volume via infra; not good
• BullMQ would add easy ability to add a button for cleaning jobs in UI
• Also would offer transparency: easy to show the currently running/queued jobs in UI Updater view, just like backend refresh jobs are done now
• The redis "total scheduled" which doesnt even work correctly, could be removed
• Means that nats-deployment could be removed as redis is already used in updater anyways. However, I'm not sure how it works in development mode where backend uses the same redis. You could separate the redises, or they might work through the same redis without problems
• Code example is already in backend