Problems / Real world, hard
One viral link gets most of the reads. The cache in front of the store hits only 60%, and its single shard handles at most 1,500 operations a second.
Free to open. Runs in your browser, no sign-up.
Success ≥ 99%, p99 ≤ 300 ms.
4,000 requests a second for 30 seconds: 95% redirect (reads), 5% shorten (writes).
What each component does in this design, and what calls it. The simulator mimics each one from these settings.
| Component | Its job here | Settings |
|---|---|---|
| Load balancerLoad balancer | Traffic enters here.Calls API.Layer 4 balancers route by IP and port; layer 7 balancers read HTTP and can route by path or header. | – |
| APIApp service | Load balancer forwards to it.Calls Redis cache (checked first), Key-value store.Little's law: throughput ≈ concurrency ÷ latency. 200 slots at 50 ms serve about 4,000 req/s. | 4 replicas × 200 slots · timeout 300 ms |
| Redis cacheCache | API checks it first for reads.An in-memory cache like Redis usually answers in well under a millisecond inside one region. | 60% hit rate |
| Key-value storeKey-value store | API reads it when Redis cache misses. API calls it for writes.Key-value stores spread keys over partitions; one very popular key can overload a single partition. | 4 partitions × 400 ops/s |
Try it first. Open one hint at a time if you get stuck.
Two cache limits at once: the hit rate and the operations a second.
Rejected cache calls and misses both land on the hot partition.
Raise the hit rate for the popular keys, and add a shard.
Raise the cache hit rate ("cache hit rate 95%") and add shards ("cache max ops 6000").