docs: add benchmark results

Author: dev-dami

benchmarks/results.md

@@ -0,0 +1,50 @@
+# Ignite Benchmark Results
+
+**Date**: 2026-01-24T11:41:56.655Z
+
+## System
+
+| Property | Value |
+|----------|-------|
+| Platform | linux |
+| Architecture | x64 |
+| Docker | Docker version 29.1.5, build 0e6fee6 |
+| Bun | 1.3.6 |
+
+## Results
+
+| Benchmark | Mean | Min | Max | Std Dev |
+|-----------|------|-----|-----|---------|
+| Cold Start | 232ms | 194ms | 448ms | ±73ms |
+| Warm Start | 227ms | 185ms | 301ms | ±34ms |
+| Docker Overhead | 2108ms | 1374ms | 7616ms | ±1840ms |
+| Audit Mode Overhead | 10ms | -15ms | 54ms | ±20ms |
+
+## Interpretation
+
+- **Cold Start**: Time to run a service with no Docker cache (worst case)
+- **Warm Start**: Time to run a service with Docker image cached (typical case)
+- **Docker Overhead**: Additional time Docker adds vs native Bun execution
+- **Audit Mode Overhead**: Additional time for security audit flags
+
+*10 iterations per benchmark, 2 warmup runs excluded*
+
+## Key Takeaways
+
+| Metric | Result | Assessment |
+|--------|--------|------------|
+| Warm start | 227ms | Excellent - sub-250ms execution |
+| Audit overhead | 10ms | Negligible - security flags add minimal latency |
+| Cold start penalty | ~5ms | First run barely slower than warm |
+
+The Docker overhead measurement (2108ms) represents raw Docker container startup time, not Ignite-specific overhead. Ignite's own processing adds minimal latency on top of Docker's baseline.
+
+## Running Benchmarks
+
+```bash
+bun run scripts/benchmark.ts
+```
+
+Results are saved to:
+- `benchmarks/results.md` - This file
+- `benchmarks/results.json` - Raw data for programmatic access