performance-profiler skill

Overview

This skill profiles application performance to find and remove bottlenecks across CPU, memory, I/O, and database layers. It guides a structured workflow from establishing a baseline to verifying improvements, helping teams produce faster, more reliable software. Practical steps cover both application-level hotspots and database query optimization for end-to-end speed gains.

How this skill works

The skill runs targeted profilers and diagnostics to collect metrics for CPU, memory, and I/O usage, plus database query execution details. It analyzes hotspots and slow queries, recommends optimizations such as code changes, index adjustments, or connection pool tuning, and then reruns measurements to verify impact. Commands and workflows are organized so you can repeat profiling, optimize, and verify results reliably.

When to use it

• Before major releases to ensure performance regressions are not introduced
• When users report slow responses or high latency in production
• During load tests to identify scaling limits and hotspots
• When database queries or transaction times spike
• When reducing infrastructure costs by improving efficiency

Best practices

• Establish a clear baseline of performance metrics before changes
• Profile in environments that mirror production for realistic results
• Target the highest-impact hotspots first using Pareto principles
• Combine application and database profiling for complete root-cause analysis
• Measure after every change to confirm actual improvement

Example use cases

• Detecting a CPU-bound loop causing request queue buildup and fixing it to reduce latency
• Profiling memory leaks during long-running jobs and implementing fixes to lower memory usage
• Analyzing slow SQL queries, adding appropriate indexes, and rewriting queries to cut response times
• Tuning connection pooling to prevent database contention under peak load
• Validating that a caching layer delivers expected throughput improvements after deployment

FAQ