performance-profiler

What This Skill Does

The Performance Profiler skill is a robust, engineering-focused tool designed for systematic identification and resolution of performance bottlenecks in Node.js, Python, and Go applications. It functions as an analytical engine that provides deep visibility into CPU usage, memory consumption, and database performance. By integrating tools like flamegraphs, py-spy, pprof, and load testing frameworks such as k6 and Artillery, the skill enables developers to move from guesswork to evidence-based optimization. It enforces a strict 'measure first' philosophy, ensuring that every architectural or code-level change is validated against a established baseline to guarantee meaningful improvements in P99 latency, throughput, and error rates.

Installation

To integrate this skill into your environment, run the following command in your terminal or within the ClawHub interface: clawhub install openclaw/skills/skills/alirezarezvani/cs-performance-profiler

Use Cases

• Latency Bottleneck Analysis: Pinpointing exactly why P99 latency exceeds SLAs under load.
• Memory Leak Detection: Analyzing heap snapshots to identify persistent objects causing memory growth over time.
• Database Performance Tuning: Identifying N+1 query patterns and slow SQL execution via EXPLAIN ANALYZE.
• Bundle Optimization: Auditing frontend bundle sizes using webpack or Next.js analyzers to improve initial load times.
• Capacity Planning: Running simulated traffic spikes using k6 or Artillery to determine the breaking point of your infrastructure.

Example Prompts

1. "I am seeing 3-second P99 latencies on the /api/orders endpoint. Can you help me profile the service to find the bottleneck?"
2. "My Node.js container's memory usage grows linearly and triggers OOM kills every 4 hours. Can you assist in taking a heap snapshot and identifying the leak?"
3. "Run a load test with 100 virtual users against our staging endpoint and provide a baseline report for latency and error rate."

Tips & Limitations

Always ensure you have a dedicated staging environment that mirrors production configurations before running heavy load tests, as these can impact database state or external API limits. The skill is most effective when access to monitoring tools like New Relic, Datadog, or local profiler outputs (e.g., .cpuprofile) is provided. Avoid running intensive profiling in production environments without careful throttling to prevent service disruption.