foxhunt/services/api_gateway/tests/proxy_latency_test.rs

//! API Gateway Proxy Latency Test
//!
//! Quick validation test for proxy latency against <1ms target
//! Measures REAL gRPC calls: API Gateway (50051) → Trading Service (50052)

#[path = "common/mod.rs"]
mod common;

use anyhow::Result;
use std::time::Instant;
use tonic::{metadata::MetadataValue, Request};
use uuid::Uuid;

use api_gateway::foxhunt::tli::{
    trading_service_client::TradingServiceClient, OrderSide, OrderType, SubmitOrderRequest,
};

/// Create authenticated request with JWT
fn create_test_request() -> Request<SubmitOrderRequest> {
    let (token, _jti) = common::generate_test_token(
        "bench-user",
        vec!["trader".to_string()],
        vec!["trading.submit_order".to_string()],
        3600,
    )
    .unwrap();

    let order = SubmitOrderRequest {
        symbol: "BTC/USD".to_string(),
        side: OrderSide::Buy as i32,
        order_type: OrderType::Limit as i32,
        quantity: 1.0,
        price: Some(50000.0),
        stop_price: None,
        time_in_force: "GTC".to_string(),
        client_order_id: Uuid::new_v4().to_string(),
    };

    let mut request = Request::new(order);
    let auth_value = MetadataValue::try_from(format!("Bearer {}", token)).unwrap();
    request.metadata_mut().insert("authorization", auth_value);

    request
}

#[tokio::test]
#[ignore = "Run manually: cargo test -p api_gateway proxy_latency --ignored -- --nocapture"]
async fn test_proxy_cold_start_latency() -> Result<()> {
    println!("\n=== Test 1: Cold Start Latency ===");

    let mut latencies = Vec::new();

    // Measure 10 cold starts
    for i in 0..10 {
        let start = Instant::now();

        let mut client = TradingServiceClient::connect("http://localhost:50051").await?;
        let request = create_test_request();

        let _ = client.submit_order(request).await;

        let elapsed = start.elapsed();
        latencies.push(elapsed);

        println!("  Cold start {}: {:?}", i + 1, elapsed);
    }

    latencies.sort();
    let median = latencies[latencies.len() / 2];
    let p99 = latencies[(latencies.len() as f64 * 0.99) as usize];

    println!("\n  📊 Cold Start Statistics:");
    println!("    Median: {:?}", median);
    println!("    P99: {:?}", p99);
    println!("    Target: <10ms (cold start allowance)");

    assert!(
        p99.as_millis() < 10,
        "P99 cold start latency {} ms exceeds 10ms target",
        p99.as_millis()
    );

    Ok(())
}

#[tokio::test]
#[ignore = "Run manually: cargo test -p api_gateway proxy_latency --ignored -- --nocapture"]
async fn test_proxy_warm_cache_latency() -> Result<()> {
    println!("\n=== Test 2: Warm Cache Latency ===");

    // Setup persistent connection
    let mut client = TradingServiceClient::connect("http://localhost:50051").await?;

    // Warmup: 100 requests
    println!("  Warming up with 100 requests...");
    for _ in 0..100 {
        let request = create_test_request();
        let _ = client.submit_order(request).await;
    }

    // Measure 1000 warm requests
    let mut latencies = Vec::new();
    println!("  Measuring 1000 warm requests...");

    for _ in 0..1000 {
        let start = Instant::now();

        let request = create_test_request();
        let _ = client.submit_order(request).await;

        latencies.push(start.elapsed());
    }

    latencies.sort();
    let p50 = latencies[latencies.len() / 2];
    let p95 = latencies[(latencies.len() as f64 * 0.95) as usize];
    let p99 = latencies[(latencies.len() as f64 * 0.99) as usize];
    let min = latencies[0];
    let max = latencies[latencies.len() - 1];

    println!("\n  📊 Warm Cache Statistics:");
    println!("    Min:    {:>8} μs", min.as_micros());
    println!("    P50:    {:>8} μs", p50.as_micros());
    println!("    P95:    {:>8} μs", p95.as_micros());
    println!("    P99:    {:>8} μs", p99.as_micros());
    println!("    Max:    {:>8} μs", max.as_micros());
    println!("    Target: < 1,000 μs (1ms)");
    println!("\n  Wave 132 Baseline: 21-488μs warm");

    if p99.as_micros() < 1000 {
        println!("  ✅ PASS: P99 {} μs < 1ms target", p99.as_micros());
    } else {
        println!("  ❌ FAIL: P99 {} μs >= 1ms target", p99.as_micros());
    }

    assert!(
        p99.as_micros() < 1000,
        "P99 latency {} μs exceeds 1ms target",
        p99.as_micros()
    );

    Ok(())
}

#[tokio::test]
#[ignore = "Run manually: cargo test -p api_gateway proxy_latency --ignored -- --nocapture"]
async fn test_proxy_overhead_comparison() -> Result<()> {
    println!("\n=== Test 3: Proxy Overhead (Proxied vs Direct) ===");

    // Setup both clients
    let mut proxy_client = TradingServiceClient::connect("http://localhost:50051").await?;
    let mut direct_client = TradingServiceClient::connect("http://localhost:50052").await?;

    // Warmup both
    println!("  Warming up proxy and direct clients...");
    for _ in 0..100 {
        let r1 = create_test_request();
        let r2 = create_test_request();
        let _ = proxy_client.submit_order(r1).await;
        let _ = direct_client.submit_order(r2).await;
    }

    // Measure proxy latency
    let mut proxy_latencies = Vec::new();
    for _ in 0..1000 {
        let start = Instant::now();
        let request = create_test_request();
        let _ = proxy_client.submit_order(request).await;
        proxy_latencies.push(start.elapsed());
    }

    // Measure direct latency
    let mut direct_latencies = Vec::new();
    for _ in 0..1000 {
        let start = Instant::now();
        let request = create_test_request();
        let _ = direct_client.submit_order(request).await;
        direct_latencies.push(start.elapsed());
    }

    proxy_latencies.sort();
    direct_latencies.sort();

    let proxy_p50 = proxy_latencies[proxy_latencies.len() / 2];
    let direct_p50 = direct_latencies[direct_latencies.len() / 2];
    let proxy_p99 = proxy_latencies[(proxy_latencies.len() as f64 * 0.99) as usize];
    let direct_p99 = direct_latencies[(direct_latencies.len() as f64 * 0.99) as usize];

    let overhead_p50 = proxy_p50.saturating_sub(direct_p50);
    let overhead_p99 = proxy_p99.saturating_sub(direct_p99);

    let overhead_percent_p50 = if direct_p50.as_micros() > 0 {
        ((overhead_p50.as_micros() as f64 / direct_p50.as_micros() as f64) * 100.0) as u32
    } else {
        0
    };

    println!("\n  📊 Proxy vs Direct Comparison:");
    println!("    Direct P50:   {:>8} μs", direct_p50.as_micros());
    println!("    Proxy P50:    {:>8} μs", proxy_p50.as_micros());
    println!(
        "    Overhead P50: {:>8} μs ({}%)",
        overhead_p50.as_micros(),
        overhead_percent_p50
    );
    println!();
    println!("    Direct P99:   {:>8} μs", direct_p99.as_micros());
    println!("    Proxy P99:    {:>8} μs", proxy_p99.as_micros());
    println!("    Overhead P99: {:>8} μs", overhead_p99.as_micros());
    println!("\n  Target: Proxy overhead < 100μs");

    if overhead_p99.as_micros() < 100 {
        println!(
            "  ✅ PASS: Overhead {} μs < 100μs",
            overhead_p99.as_micros()
        );
    } else {
        println!(
            "  ⚠️  WARNING: Overhead {} μs >= 100μs",
            overhead_p99.as_micros()
        );
    }

    Ok(())
}

#[tokio::test]
#[ignore = "Run manually: cargo test -p api_gateway proxy_latency --ignored -- --nocapture"]
async fn test_connection_pool_impact() -> Result<()> {
    println!("\n=== Test 4: Connection Pool Impact ===");

    for concurrency in [1, 10, 50, 100] {
        let start = Instant::now();

        let mut handles = vec![];
        for _ in 0..concurrency {
            let handle = tokio::spawn(async move {
                let mut client = TradingServiceClient::connect("http://localhost:50051")
                    .await
                    .unwrap();
                let request = create_test_request();
                let _ = client.submit_order(request).await;
            });
            handles.push(handle);
        }

        for handle in handles {
            let _ = handle.await;
        }

        let elapsed = start.elapsed();
        let avg_per_request = elapsed / concurrency;

        println!(
            "  Concurrency {:<3}: Total {:>6?}, Avg/req {:>6?}",
            concurrency, elapsed, avg_per_request
        );
    }

    println!("\n  ✅ Connection pool test complete");

    Ok(())
}