capacity-planning Skill

Goal

This skill guides the estimation of infrastructure needs based on usage patterns.

Steps

1. The Universal Formula

Required Replicas = (Target RPS * Avg Request Latency) / Max Thread Utilization

Example:

• Target: 1000 RPS
• Latency: 0.2s (200ms)
• 1 Core handles: 1 / 0.2 = 5 req/s (Sync) OR much higher if Async.
• Rule of Thumb: Run load tests to find "Breaking Point" RPS per instance.

2. Resource Dimensions