Design a Distributed Job Scheduler · System Design Interview

Design a Distributed Job SchedulerDesign the service that runs other teams’ jobs on a schedule — cron, but across a fleet and with the guarantees cron does not give.What it has to do

A team registers a job with a schedule, and it runs at those times

A job that fails is retried according to its policy

A team can see whether a given run succeeded, and why it did not

Scale

Registered jobs: 5 million

Runs per day: 400 million

Out of scope

The job’s own code and runtime

Resource quotas and billing

Workflows and dependencies between jobs

Practise this challengeAll system design challenges

LetsDiscuss

Loading your session…