performance-testing

Plan and run load, stress, and spike tests for web/API services.

364|53|Updated May 9, 2026

Search Tags:#automation #load testing #performance testing #k6 #stress testing #bottleneck analysis #latency thresholds

Ready to run

npx skills add https://github.com/cosmicstack-labs/mercury-agent-skills --skill performance-testing-cosmicstack-labs

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: performance-testing
Source: https://github.com/cosmicstack-labs/mercury-agent-skills/tree/main/categories/testing-qa/performance-testing
Command: npx skills add https://github.com/cosmicstack-labs/mercury-agent-skills --skill performance-testing-cosmicstack-labs

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Performance testing helps you verify that your system stays reliable and fast under real load, including spikes, stress, and long-running usage.

Core Features & Use Cases

Load, stress, and spike test planning: Use clear test types (smoke, load, stress, spike, endurance, soak) to cover different reliability risks.
k6-ready patterns: Define stages and enforce latency/error thresholds to prevent regressions.
Bottleneck analysis flow: Turn slow endpoints into actionable causes across database, CPU/memory, and external service latency.

Quick Start

Run a load test against your API and enforce p(95) latency and failure-rate thresholds to confirm it meets your performance targets.

Frequently Asked Questions about performance-testing

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I set up load testing to find bottlenecks in my API?▼

To find bottlenecks, you run stress testing by defining load stages and enforcing latency thresholds. This process validates API reliability under traffic spikes and iteratively investigates slow endpoints across databases and compute resources.

What is the difference between stress, spike, and soak testing?▼

Stress testing evaluates breaking points, spike testing simulates sudden traffic surges, and soak testing assesses long-running usage. These performance testing types validate system reliability under different load conditions and traffic patterns.

How do I enforce latency thresholds during k6 load testing?▼

k6 load testing enforces latency thresholds by defining p(95) latency and failure-rate targets within test scripts. This prevents performance regressions by failing test runs when response times exceed established limits.

Can I use this performance testing approach for web services and their database dependencies?▼

Yes, performance testing validates web services and their dependencies by analyzing bottlenecks across endpoints, databases, compute resources, and external services. It supports an iterative investigation workflow to resolve slow endpoints.

What's the best way to analyze database and CPU bottlenecks found during stress testing?▼

Stress testing bottleneck analysis uses an iterative investigation workflow across endpoints, databases, CPU, memory, and external service latency. This turns slow endpoints into actionable root causes for targeted system optimization.