Design a Distributed Job Scheduler (like Cron)

This lesson designs a Distributed Job Scheduler, often called "Distributed Cron": a service that calls other services at the times they asked to be called. The hard part is not storing the schedules. It is finding the jobs that are due right now among 100 million rows, every second, without losing a run when a server crashes.

The name comes from cron, the Linux tool that runs commands on a schedule. This design shares only cron's schedule format. Real cron runs on one machine, so if that machine is down at the scheduled time, the run is missed, and it has no retries or central history.