horus/benches/QUICK_START.md

Horus Benchmarks - Quick Start

1. Start the Stack

# Terminal 1: Start Redis
redis-server

# Terminal 2: Start Horus
cd /Users/timurgordon/code/git.ourworld.tf/herocode/horus
RUST_LOG=info ./target/release/horus all --admin-secret SECRET --kill-ports

2. Run Benchmarks

Option A: Use the helper script (recommended)

./benches/run_benchmarks.sh

Option B: Run directly with cargo

# All benchmarks
cargo bench

# Specific benchmark suite
cargo bench --bench horus_stack
cargo bench --bench stress_test

# Specific test
cargo bench --bench horus_stack -- supervisor_discovery

# Quick run (fewer samples)
cargo bench -- --quick

3. View Results

# Open HTML report in browser
open target/criterion/report/index.html

# Or on Linux
xdg-open target/criterion/report/index.html

Available Benchmark Suites

horus_stack - Standard Performance Tests

• API discovery and metadata
• Runner management
• Job operations
• Concurrency tests
• Health checks
• API latency measurements

stress_test - Load & Stress Tests

• High-frequency job submissions (50-200 jobs)
• Sustained load testing
• Large payload handling (1KB-100KB)
• Rapid API calls (100 calls/test)
• Mixed workload scenarios
• Connection pool exhaustion (10-100 clients)

memory_usage - Memory Profiling

• Job object memory footprint (10-200 jobs)
• Client instance memory overhead (1-100 clients)
• Payload size impact on memory (1KB-1MB)
• Memory growth patterns under load

Common Commands

# Run only fast benchmarks
cargo bench -- --quick

# Save baseline for comparison
cargo bench -- --save-baseline main

# Compare against baseline
cargo bench -- --baseline main

# Run with verbose output
cargo bench -- --verbose

# Filter by name
cargo bench -- concurrent
cargo bench -- stress

# Run specific benchmark group
cargo bench --bench horus_stack -- api_latency

# Run memory benchmarks
cargo bench --bench memory_usage

# Run memory benchmarks with verbose output (shows memory deltas)
cargo bench --bench memory_usage -- --verbose

Troubleshooting

"Connection refused"

• Make sure Horus stack is running
• Check ports: 3030 (supervisor), 8081 (osiris), 9652/9653 (coordinator)

"Job timeout"

• Increase timeout in benchmark code
• Check that runners are registered: curl http://127.0.0.1:3030 (requires POST)

Slow benchmarks

• Close other applications
• Use --quick flag for faster runs
• Reduce sample size in benchmark code

Performance Expectations

Test	Expected Time
supervisor_discovery	< 10ms
supervisor_get_info	< 5ms
job_full_lifecycle	< 100ms
concurrent_jobs (10)	< 500ms
stress_high_frequency (50)	< 2s

Next Steps

• See benches/README.md for detailed documentation
• Modify benches/horus_stack.rs to add custom tests
• Check target/criterion/ for detailed reports