System Design · Snowflake · Hard
Requirements Expose the following API: submit(job, cron_expression) pause(job_id) — executions that have already started are allowed to complete, while later executions must not begin. resume(job_id) — permits upcoming scheduled executions again. cancel(job_id) Treat worker capacity as practically unbounded. The design should prioritize correct scheduling behavior rather than worker-pool scaling. Durability: job metadata, cron schedules, and the active lifecycle state…
Checking your access…