optimizing-performance_skill

Overview

This skill analyzes and optimizes application performance across frontend, backend, and database layers. It guides you through a data-driven workflow: measure baseline, identify bottlenecks, apply targeted fixes, and verify improvements. Use it to diagnose slowness, reduce load times, and improve system efficiency.

How this skill works

The skill inspects runtime profiles, web vitals, database query plans, and resource usage to locate CPU, memory, I/O, and database bottlenecks. It recommends focused changes—bundle splitting, memoization, optimized queries, caching, queues, and algorithmic improvements—then prescribes how to re-measure and validate gains. Metrics and checks ensure optimizations are safe and repeatable.

When to use it

• Investigating slow page loads or poor Core Web Vitals
• Diagnosing high API latency, tail latency, or error spikes
• Finding and fixing inefficient database queries or locks
• Reducing frontend bundle size and render cost
• Optimizing background jobs or long-running tasks

Best practices

• Never optimize without data: profile before and after every change
• Target the highest-impact bottleneck first (CPU, I/O, DB, or memory)
• Apply multi-layer caching: in-memory, Redis, then DB
• Prefer non-blocking work: push long tasks to background queues
• Measure using the same tools and scenarios to ensure apples-to-apples comparison

Example use cases

• Use profiling to find and fix an N+1 query by converting many queries into a single JOIN
• Reduce first contentful paint by code-splitting and lazy-loading heavy components
• Cut API P95 latency by adding Redis caching and optimizing slow SQL with EXPLAIN
• Prevent server stalls by moving video processing to a background queue
• Replace an O(n²) algorithm with an O(n) approach to eliminate CPU hotspots