What problem does it solve?
This Skill helps you reason about latency, throughput, queue depth, and utilization with a concrete queueing model instead of intuition alone. It is useful when systems slow down under load, tail latency spikes, or scaling decisions around workers, threads, and pools are being made without a clear explanation of what will actually improve performance.
Core Features & Use Cases
- Little's Law analysis: Relates in-flight work, arrival rate, and time in system so you can derive the missing quantity from measurements you already have.
- Utilization and bottleneck diagnosis: Identifies whether latency is driven by high utilization, service-time variance, bounded queues, or the wrong bottleneck being optimized.
- Capacity planning and pool sizing: Supports thread pool sizing, connection limits, autoscaling reviews, timeout decisions, and deciding whether to add concurrency or reduce service time.
- Use case: When an API's p99 latency suddenly worsens after traffic grows, use this Skill to compute the operating point, locate the system on the utilization curve, and determine whether the right action is adding workers, reducing service time, splitting workloads, or fixing a downstream bottleneck.
Quick Start
Ask the littles-law-reasoning skill to analyze your service using current request rate, average latency, in-flight count, worker count, and estimated service time.
