Problems / Design a product, medium
Design a cron-as-a-service from an empty canvas: millions of scheduled jobs fired on time, each run at least once and its effect exactly once.
Unlock it once for 15 credits; everything inside it is free (runs, changes, Submit, results, hints). Sign in with Google to start; new accounts get free credits.
Request types to route: status (60%, read), create (20%, write), tick (20%, write). Budget: $4,500 a month.
You start from an empty canvas. When you run the tests, each one simulates your design with its own traffic and failures:
| Test | What it does | Passes when |
|---|---|---|
| Normal day | The normal day: the traffic in the brief, nothing broken. While this test runs, brokers redeliver unacknowledged messages after 5 s and 0.2% of acknowledgements are lost, as in production: only idempotent consumers avoid duplicates. | Success ≥ 99.9%; p99 latency ≤ 200 ms (status); p99 latency ≤ 300 ms (create); No duplicate side effects |
| Top of the hour | At 5 s the top of the hour arrives: every hourly job is due at once and traffic triples for 8 s. | Success ≥ 99%; p99 latency ≤ 500 ms (tick) |
| A server dies | At 4 s one server of the first app service on the request path dies for 10 s. | Success ≥ 99% |
| Workers stall mid-job | At 4 s job runs take 15× longer for 8 s (a slow downstream, long GC pauses): runs outlive their lease and the queue delivers them again. While this test runs, brokers redeliver unacknowledged messages after 5 s and 0.2% of acknowledgements are lost, as in production: only idempotent consumers avoid duplicates. | Success ≥ 99%; No duplicate side effects |
| Budget: $4,500/month | The normal day's design must cost at most $4,500 a month. | Cost ≤ $4,500/month |