Allspaw frames capacity planning as a simple repeating loop, not a one-time calculation. It starts from one question: what performance do you need from your site? Then four steps: 1. **Measure** how well the current infrastructure works — characterize each component's workload against your performance requirements. 2. **Predict** what you'll need in future — project from past performance, then marry the prediction to what you can afford and a realistic timeline. 3. **Deploy** the new capacity with proven tools and techniques. 4. **Iterate** — calibrate the plan over time ("rinse, repeat"). The goal sits between two failures: too little hardware (irate users) and too much (wasted money) — the supermarket manager scheduling cashiers against a variable, growing flow of customers. The cycle is architecture-specific: the calculations depend on your stack. --- *Source: [[The Art of Capacity Planning]] (John Allspaw, O'Reilly 2008) — Ch 1 — Goals, Issues, and Processes in Capacity Planning*