congiuluc/performance-backend

Backend Performance Optimization (.NET)

When to Use

• Adding caching to frequently accessed data
• Optimizing database queries or repository methods
• Implementing pagination for large datasets
• Adding response compression to the API
• Profiling and fixing backend performance bottlenecks

Official Documentation

• Response Caching in ASP.NET Core
• Output Caching
• Response Compression
• EF Core Performance

Procedure

1. Identify bottleneck: database query, I/O wait, serialization, missing index
2. Apply caching and compression patterns
3. Review caching sample
4. Add pagination for list endpoints (offset or cursor-based)
5. Use AsNoTracking() for read-only EF Core queries
6. Project only needed columns with .Select()
7. Measure before and after with structured logging
8. Add performance-sensitive tests for critical paths

Performance Review Checklist

Use this checklist when reviewing backend code for performance issues:

| # | Check | Tool / Method | |---|-------|---------------| | 1 | All list endpoints have pagination | Code review | | 2 | Read-only queries use AsNoTracking() | Code review | | 3 | Only needed columns projected with .Select() | Code review | | 4 | No N+1 query patterns (use .Include() or batch) | EF Core logging | | 5 | Response compression enabled (Brotli + Gzip) | curl -H 'Accept-Encoding: br' | | 6 | Hot data cached with appropriate TTL | Code review | | 7 | Cache invalidation on write operations | Code review | | 8 | Async/await used for all I/O (no .Result or .Wait()) | Code review | | 9 | CancellationToken propagated through call chain | Code review | | 10 | No unbounded ToListAsync() without Take() | Code review | | 11 | Database indexes on frequently queried columns | Schema review | | 12 | Connection pooling configured correctly | Config review |

Profiling Workflow

1. Enable EF Core query logging: optionsBuilder.LogTo(Console.WriteLine)
2. Use dotnet-counters for real-time metrics: dotnet-counters monitor
3. Use dotnet-trace for detailed profiling: dotnet-trace collect
4. Check Aspire dashboard for request timing and dependencies
5. Run BenchmarkDotNet for micro-benchmarks on critical paths
6. Compare before/after with structured Serilog timing logs

Official Profiling References

• dotnet-counters
• dotnet-trace
• BenchmarkDotNet
• EF Core Query Logging