Dev Performance Bottleneck Analysis
===================================

Create a performance bottleneck analysis framework for [application type].

Application details:
- Type: [web app/API/mobile/data pipeline]
- Tech stack: [list]
- Symptoms: [slow page load/high latency/memory leak/high CPU]
- Environment: [production/staging]
- Scale: [requests per second/users]

PERFORMANCE ANALYSIS FRAMEWORK

1. IDENTIFY THE BOTTLENECK

   MEASUREMENT FIRST
   - Profile before optimizing
   - Tools: [Chrome DevTools/New Relic/DataDog/py-spy]
   - Establish baseline metrics:
     * Response time: [Xms]
     * Throughput: [requests/sec]
     * Error rate: [%]
     * Resource usage: [CPU%/Memory MB]

2. COMMON BOTTLENECK CATEGORIES

   DATABASE
   - Slow queries: [use EXPLAIN ANALYZE]
   - Missing indexes: [check query plans]
   - N+1 queries: [check ORM generated SQL]
   - Connection pool exhaustion: [check pool settings]
   - Lock contention: [check pg_locks/slow query log]

   APPLICATION CODE
   - Inefficient algorithms: [O(n²) where O(n) possible]
   - Unnecessary loops: [check nested iterations]
   - Memory leaks: [check heap snapshots]
   - Synchronous blocking: [async opportunities]
   - Serialization overhead: [JSON/XML processing]

   NETWORK
   - Too many HTTP requests: [bundle/cache]
   - Large payload sizes: [compress/paginate]
   - No CDN for static assets
   - DNS resolution: [latency]

   INFRASTRUCTURE
   - CPU bound: [vertical scale or optimize]
   - Memory bound: [increase RAM or fix leaks]
   - I/O bound: [SSD/caching/async I/O]
   - Cache hit rate: [target 80%+]

3. OPTIMIZATION PRIORITY
   - Fix highest impact first
   - Measure after each change
   - Do not optimize prematurely
   - Document what you changed and why

4. SOLUTIONS BY CATEGORY

   DATABASE FIXES
   - Add index: [on frequently queried columns]
   - Query optimization: [rewrite slow queries]
   - Caching: [Redis/Memcached for frequent reads]
   - Read replicas: [for read-heavy workloads]
   - Connection pooling: [PgBouncer/HikariCP]

   CODE FIXES
   - Caching results: [memoization]
   - Lazy loading: [load only what needed]
   - Pagination: [limit result sets]
   - Async processing: [background jobs]
   - Algorithm improvement: [better complexity]

5. MONITORING AFTER FIX
   - Set up alerts for regression
   - Dashboard for key metrics
   - Load test before deploying

Source: https://promptzyo.com/prompt/dev-performance-bottleneck-analysis