PatternKit Benchmarks

This project contains reportable BenchmarkDotNet comparisons for PatternKit fluent builders and source-generated factories.

Run the full benchmark suite from the repository root:

dotnet run -c Release --framework net10.0 --project benchmarks/PatternKit.Benchmarks -- --artifacts artifacts/benchmarks --join

Run one pattern family:

dotnet run -c Release --framework net10.0 --project benchmarks/PatternKit.Benchmarks -- --filter *LeaderElection* --artifacts artifacts/benchmarks --join

Every pattern-specific scenario benchmark class must:

• Compare fluent and source-generated routes in the same benchmark class.
• Split construction/setup overhead from execution overhead where the pattern has meaningful runtime work.
• Keep realistic domain names and payloads so benchmark results map back to production examples.
• Emit memory diagnostics and markdown, CSV, and JSON reports through the shared benchmark config.
• Use BenchmarkDotNet categories for the pattern family, pattern name, route, and benchmark phase.

Coverage matrix benchmarks under Coverage/ include every pattern from PatternKitPatternCatalog and every *Generator.cs source file from src/PatternKit.Generators. The TinyBDD production-readiness tests fail when a catalog pattern or generator is missing from that matrix. Pattern-specific benchmarks can then add deeper scenario timing while the matrix keeps top-to-bottom coverage complete.

Published benchmark snapshots live in docs/guides/benchmarks.md so users can compare fluent and source-generated timing, overhead, and allocation without running the suite first.