Scale, Performance, and Consistency

Principle 3: Architect for Scalability

Scalability is the ability to handle a changing workload without unacceptable loss of performance or control over cost.

A growing dataset may stress storage, while a burst of concurrent jobs stresses processing capacity. Those pressures can require different changes.

Find the limiting resource and examine how the workload uses it. Adding workers will not fix a component that remains a shared bottleneck.

Identify what is growing: records, requests, concurrent jobs, or state. Add capacity where that growth creates a bottleneck rather than scaling every component equally.