Database Queries Per Second
What you'll learn: How to calculate expected read and write queries per second (QPS) to determine if your database can handle a URL shortener's traffic without falling over.
Why QPS Matters
When designing systems like a URL shortener, you can't just pick a database and hope it works. Databases have throughput limits—they can only handle so many operations per second before performance degrades or requests start failing. By calculating your expected QPS before launch, you inform critical scaling decisions: single server vs replicas, sharding strategy, caching layer size, and more.
Calculating Read QPS
Using estimates from previous lessons (DAUs, requests per user):
If you expect 10 million URL redirects per day (your read-heavy traffic), spread across 24 hours with peak traffic at 3× average:
- Average reads/sec = 10,000,000 / 86,400 ≈ 116 QPS
- Peak reads/sec ≈ 116 × 3 = ~350 QPS
Calculating Write QPS
If users create 1 million new short URLs per day:
- Average writes/sec = 1,000,000 / 86,400 ≈ 12 QPS
- Peak writes/sec ≈ 12 × 3 = ~36 QPS
The Read/Write Ratio Insight
Your URL shortener has a 10:1 read-to-write ratio (10M reads vs 1M writes). This heavily influences architecture:
- A single MySQL instance handles 1,000–5,000 QPS easily → you're safe for now
- But plan read replicas or caching for growth
- Write QPS is low → single primary database sufficient initially
When to Scale
If projections show >1,000 read QPS, you'll need read replicas or aggressive caching. If >500 write QPS, consider sharding or vertical scaling.
Key Takeaway: Calculating read and write QPS from traffic estimates lets you right-size your database architecture, avoid over-engineering, and plan scaling thresholds before they become emergencies.