anujagrawal380/distributed-job-queue ? reverse-engineered prompt

Build me a durable job queue service in Go that can handle lots of background work without losing jobs if the server crashes.

I want an API where clients can submit jobs with a payload, priority, retries, and either a delay or a future run time. Workers should be able to lease jobs, complete them, or fail them, and failed jobs should retry until they hit a dead letter state. The queue should remember everything on disk so it can recover after a restart, and it should also support a cluster mode where multiple nodes stay in sync and keep working if the leader goes down.

Please add simple API key authentication with different access levels for clients, workers, and admins. I also want a live web dashboard on the home page that shows queue stats, recent jobs, and updates in real time. If possible, include a small Go SDK example and a chaos test command that can hammer the queue while restarting it to prove there’s zero loss. If you need to check current docs for any libraries, feel free to look them up online.