Agent skill · data analytics · jeremylongshore
running-performance-tests
Execute load testing, stress testing, and performance benchmarking.
Why this skill is useful
Adds specific performance testing scripts and methodologies that the AI wouldn't reliably generate on its own.
What it needs
Requires artillery, autocannon, jmeter, k6, locust installed locally. About 3k tokens when loaded. Last updated 2026-08-07. 2,604 stars on the source repository.
What this skill does
Performance Test Suite Overview Execute load testing, stress testing, and performance benchmarking to identify bottlenecks, establish baseline metrics, and verify SLA compliance. Supports k6 (recommended), Artillery, Apache JMeter, Locust (Python), and autocannon (Node.js). Prerequisites Performance testing tool installed (k6, artillery, locust, jmeter, or autocannon) Target application deployed in a production-like environment (not local dev) Baseline performance metrics or SLA targets (e.g., p95 < 200ms, 99.9% availability) Monitoring stack accessible (Grafana, CloudWatch, Datadog) for resource metrics during tests Test data sufficient to avoid cache-only responses Instructions 1. Define performance test scenarios based on production traffic patterns: Load test: Simulate expected peak traffic (e.g., 500 concurrent users for 10 minutes). Stress test: Ramp beyond expected capacity to find the breaking point. Spike test: Sudden burst of traffic (0 to 1000 users in 10 seconds). Soak test: Sustained moderate load for extended duration (1-4 hours) to detect memory leaks. 2. Create test scripts targeting critical endpoints: Identify the top 5-10 most-hit API endpoints from production access logs. Include both read (GET) and write (POST/PUT/DELETE) operations. Simulate realistic user behavior with think time between requests. Use parameterized data to avoid cache-only hits (randomize query parameters, user IDs). 3. Configure load profiles: Define virtual user (VU) ramp-up stages (e.g., 10 VUs for 1 minute, then 50 VUs for 5 minutes). Set test duration appropriate to the scenario (load: 10-15 min, soak: 1-4 hours). Configure request timeouts matching production settings. 4. Execute the performance test: Run from a machine with sufficient network bandwidth and CPU. Avoid running from the same host as the application under test. Monitor application metrics (CPU, memory, DB connections) during execution. 5. …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills jeremylongshore/running-performance-tests