Design a Delayed Job Scheduler

Design a service that accepts jobs to run at a specified future time and dispatches them when they become due. ### Clarifying Questions The reported prompt names a delayed job scheduler without a concrete workload or delivery contract. Clarify scheduling precision, job volume, payload size, retry policy, cancellation, and whether duplicate execution is acceptable. For the worked architecture, choose durable schedules and at-least-once dispatch with idempotent execution. These are explicit design choices rather than reported requirements. ### What a Strong Answer Covers A schedule data model, due-time indexing, concurrent claiming, dispatch and execution boundaries, crash recovery, retries, time handling, and operational visibility. Explain how the system handles a new earlier job while a dispatcher waits. ### Follow-up Questions - How would cancellation race with a dispatcher that already claimed a job? - What can be guaranteed if a worker performs an external side effect and then crashes before acknowledging it? - When would a timing wheel or partitioned due-time queue be appropriate?