Can it help diagnose existing production bottlenecks?

Yes, by measuring wall-clock time at every component boundary, the skill identifies whether the slowdown is occurring at the network, cache, database, or compute level.

Why does this skill focus on p99 latency instead of averages?

Optimizing for the average (p50) often ignores the worst-case experiences that drive users away. Focusing on p99 ensures that even the slowest requests are optimized for a reliable user experience.

Which queueing models are supported?

The skill applies standard queueing theory models including M/M/1 for simple resources and M/M/c for multi-server environments to provide grounded predictions for system behavior.

How does this skill help with capacity planning?

It uses mathematical models like Little's Law to predict how your system's latency and throughput will respond as load increases, helping you identify hardware needs before they become critical.

Performance Modeling

Name: Performance Modeling
Author: sethdford

bysethdford

•

분석 및 모니터링

Models system performance and predicts latency under load to identify architectural bottlenecks using queueing theory.

The Performance Modeling skill enables developers and architects to rigorously analyze and predict system behavior under various load profiles. Grounded in queueing theory and Little's Law, it helps establish performance baselines, identify specific breaking points in infrastructure (such as database connection limits or CPU saturation), and model the impact of optimizations like read replicas or CDNs. By focusing on tail latency (p99) rather than simple averages, this skill ensures that systems remain responsive for all users during capacity planning or performance tuning phases.

주요 기능

01Generates latency-versus-load curves to pinpoint precise system breaking points.

02Identifies architectural bottlenecks across databases, caches, network, and CPU.

035 GitHub stars

04Mitigates common anti-patterns like optimizing for average latency instead of p99 tails.

05Evaluates the ROI of performance optimizations versus implementation effort.

06Predicts system latency and throughput using mathematical queueing models (M/M/1, M/M/c).

사용 사례

01Modeling the performance benefits of introducing caching layers or database read replicas.

02Diagnosing root causes of intermittent p99 latency spikes in distributed systems.

03Capacity planning for high-traffic events to predict when current infrastructure will fail.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add sethdford/claude-skills performance-modeling

For use in Claude.ai and ChatGPT

주요 기능

01Generates latency-versus-load curves to pinpoint precise system breaking points.

02Identifies architectural bottlenecks across databases, caches, network, and CPU.

035 GitHub stars

04Mitigates common anti-patterns like optimizing for average latency instead of p99 tails.

05Evaluates the ROI of performance optimizations versus implementation effort.

06Predicts system latency and throughput using mathematical queueing models (M/M/1, M/M/c).

사용 사례

01Modeling the performance benefits of introducing caching layers or database read replicas.

02Diagnosing root causes of intermittent p99 latency spikes in distributed systems.

03Capacity planning for high-traffic events to predict when current infrastructure will fail.