System-design capacity calculator

Make traffic, peak-load and storage assumptions visible before choosing an architecture.

Daily active users1–1,000,000,000Requests per user per day1–100,000Peak / average multiplier1–100Requests creating a stored record (%)0–100Bytes per stored record1–10,000,000Retention days1–3,650Total stored copies, including replicas1–10

23.15 average requests / second

115.74 estimated peak requests / second

200,000 new records / day

17.17 GiB stored data across all copies

Show the calculation and assumptions

Daily requests = users × requests per user. Average RPS = daily requests ÷ 86,400. Peak RPS = average × peak multiplier. Storage = daily requests × write percentage × bytes per new record × retention days × copies.

The model assumes each write creates one equally sized retained record. It excludes indexes, backups, compression, update reuse and network overhead. Measure those separately and test burst behavior before provisioning a service.

Turn the result into understanding

Write down why each assumption fits the product. Try doubling the traffic and retention period. Explain which component becomes constrained first, how it degrades, and which production measurement would validate the estimate.

Continue with related practice →Read worked guides

Reference: Official documentation