Zig HTTP Sentinel

A robust, thread-safe HTTP client library for Zig 0.16.0, battle-tested in production environments.

Note: This library was extracted from production high-frequency trading systems where reliability and performance are critical.

Developed by QUANTUM ENCODING LTD
Contact: [email protected]

Features

• Production-Grade API: Enterprise-level interface for all HTTP operations
• Thread-Safe: Designed for concurrent use (each thread should use its own client instance)
• Memory-Safe: Automatic memory management with proper cleanup
• Full HTTP Support: GET, POST, PUT, PATCH, DELETE, HEAD methods
• Configurable: Customizable request options including timeouts and body size limits
• Production-Ready: Extensively tested under high-load conditions

Optional Advanced Modules

• Retry Engine: Battle-tested resilience patterns including exponential backoff, circuit breakers, and adaptive retry strategies
• Connection Pool: Enterprise-grade connection pooling with health monitoring, load balancing, and multi-host support

Installation

Add this library to your build.zig.zon:

.dependencies = .{
    .http_sentinel = .{
        .url = "https://github.com/YOUR_USERNAME/zig-http-sentinel/archive/refs/tags/v1.0.0.tar.gz",
        .hash = "YOUR_HASH_HERE",
    },
},

Then in your build.zig:

const http_sentinel = b.dependency("http_sentinel", .{
    .target = target,
    .optimize = optimize,
});
exe.root_module.addImport("http-sentinel", http_sentinel.module("http-sentinel"));

Production Deployment

const std = @import("std");
const HttpClient = @import("http-sentinel").HttpClient;

pub fn main() !void {
    var gpa = std.heap.GeneralPurposeAllocator(.{}){};
    defer _ = gpa.deinit();
    const allocator = gpa.allocator();

    // Create client
    var client = HttpClient.init(allocator);
    defer client.deinit();

    // Make a GET request
    const headers = [_]std.http.Header{
        .{ .name = "Accept", .value = "application/json" },
    };
    
    var response = try client.get("https://api.example.com/data", &headers);
    defer response.deinit();

    std.debug.print("Status: {}\n", .{response.status});
    std.debug.print("Body: {s}\n", .{response.body});
}

API Reference

Enterprise Client Initialization

var client = HttpClient.init(allocator);
defer client.deinit();

HTTP Methods

All methods return a Response struct that must be deinitialized:

pub const Response = struct {
    status: http.Status,
    body: []u8,
    allocator: std.mem.Allocator,
    
    pub fn deinit(self: *Response) void;
};

GET Request

var response = try client.get(url, headers);
defer response.deinit();

POST Request

var response = try client.post(url, headers, body);
defer response.deinit();

PUT Request

var response = try client.put(url, headers, body);
defer response.deinit();

PATCH Request

var response = try client.patch(url, headers, body);
defer response.deinit();

DELETE Request

var response = try client.delete(url, headers);
defer response.deinit();

HEAD Request

var response = try client.head(url, headers);
defer response.deinit();

Advanced Options

For more control, use the WithOptions variants:

const options = HttpClient.RequestOptions{
    .max_body_size = 50 * 1024 * 1024, // 50MB
    .timeout_ns = 30 * std.time.ns_per_s, // 30 seconds
};

var response = try client.getWithOptions(url, headers, options);
defer response.deinit();

Advanced Optional Modules

Retry Engine

The retry module provides enterprise resilience patterns extracted from production HFT systems:

const RetryEngine = @import("http-sentinel/retry");

// Configure retry strategy
const config = RetryEngine.RetryConfig{
    .max_attempts = 5,
    .initial_delay_ms = 100,
    .max_delay_ms = 10000,
    .backoff_multiplier = 2.0,
    .jitter_factor = 0.1,
};

var engine = RetryEngine.init(allocator, config);
defer engine.deinit();

// Use with HTTP requests
var attempt: u32 = 0;
while (attempt < config.max_attempts) : (attempt += 1) {
    var response = client.get(url, headers) catch |err| {
        const delay = engine.calculateDelay(attempt);
        std.time.sleep(delay * std.time.ns_per_ms);
        continue;
    };
    defer response.deinit();
    break; // Success
}

Patterns Included:

• Exponential backoff with jitter
• Circuit breaker pattern
• Adaptive retry based on health scores
• Configurable retry policies

Connection Pool

Enterprise-grade connection pooling for high-throughput scenarios:

const ConnectionPool = @import("http-sentinel/pool");

// Configure pool
const config = ConnectionPool.PoolConfig{
    .max_connections = 20,
    .max_idle_connections = 10,
    .connection_timeout_ms = 5000,
    .idle_timeout_ms = 30000,
    .max_connection_lifetime_ms = 300000,
};

var pool = try ConnectionPool.init(allocator, config);
defer pool.deinit();

// Acquire and use connections
const conn = try pool.acquire("api.example.com", 443);
defer pool.release(conn);

// Use connection for HTTP operations
var response = try conn.request(.GET, "/data", headers);
defer response.deinit();

Features:

• Multi-host connection management
• Connection health monitoring
• Load balancing across backends
• Automatic connection lifecycle management
• Per-host statistics and metrics

Run the examples to see these patterns in action:

zig build retry-demo     # Demonstrates retry patterns
zig build pool-demo      # Shows connection pooling

Thread Safety & Concurrency

CRITICAL: HTTP Sentinel uses the client-per-worker pattern for concurrent operations. This is the only reliable way to do concurrent HTTP requests in Zig 0.16.0.

✅ Correct Pattern (Client-Per-Worker)

fn workerThread(allocator: std.mem.Allocator) void {
    // Each thread creates its own client
    var client = HttpClient.init(allocator);
    defer client.deinit();
    
    // Use the client safely for this worker's requests
}

❌ Wrong Pattern (Will Segfault)

// DON'T DO THIS - Shared client will segfault under load
var shared_client = HttpClient.init(allocator);
var mutex = std.Thread.Mutex{};

fn workerThread(client: *HttpClient, mutex: *std.Thread.Mutex) void {
    mutex.lock();
    defer mutex.unlock();
    // This WILL segfault even with mutex protection!
    const response = client.get(...);
}

Why: Zig 0.16.0's http.Client has internal state that is not thread-safe. The client-per-worker pattern avoids all concurrency issues.

See CONCURRENCY_PATTERN.md for detailed explanation and benchmarks.

Examples

See the examples/ directory for complete working examples including:

• Basic GET/POST requests
• JSON API interactions
• Concurrent request handling
• Error handling patterns

Run examples:

zig build examples

Example: High-Performance AI Client

Demonstrating zig-http-sentinel's enterprise capabilities with Anthropic's Claude API - proving universal applicability beyond financial systems.

Setup

1. Get your Anthropic API key from https://console.anthropic.com/
2. Set your environment variable:

   export ANTHROPIC_API_KEY=your_api_key_here
   

Run the AI Client Demo

cd examples
zig run anthropic_client.zig --deps http-sentinel

Features Demonstrated

• Production JSON Construction: Enterprise-grade payload building with proper escaping
• Professional Header Management: Complete API authentication and versioning
• Multi-Turn Conversations: Stateful conversation handling
• Robust Error Handling: Comprehensive API error processing
• Memory Safety: Proper allocation and cleanup patterns
• Performance Metrics: Token usage tracking and optimization

Sample Output

=== Zig HTTP Sentinel: High-Performance AI Client ===

🚀 Initializing high-performance AI client...
📡 Using zig-http-sentinel for enterprise-grade HTTP operations

📝 Demo 1: Production Message Processing
🤖 Claude (claude-3-haiku-20240307):
Zig excels for HTTP clients through zero-cost abstractions, compile-time safety, 
manual memory management, and cross-platform compatibility. Its performance matches 
C while preventing common networking bugs through strong typing.
📊 Tokens: 23 in, 50 out

💬 Demo 2: Multi-Turn Conversation
🤖 Claude (claude-3-haiku-20240307):
zig-http-sentinel achieves these through Zig's allocator patterns for memory efficiency,
built-in thread safety, comprehensive error types, connection pooling architecture,
and clean generic interfaces that maintain zero-cost abstractions.
📊 Tokens: 67 in, 89 out

⚡ Demo 3: Technical Analysis
🤖 Claude (claude-3-haiku-20240307):
This pattern ensures deterministic cleanup, prevents memory leaks through RAII-style
resource management, enables zero-copy optimizations, and maintains explicit control
over allocation strategies—critical for high-frequency, low-latency systems.
📊 Tokens: 124 in, 156 out

✅ All demonstrations completed successfully!
💎 zig-http-sentinel: Enterprise-grade HTTP client for production AI systems

Integration in Your Projects

const AnthropicClient = @import("your_ai_module.zig").AnthropicClient;

var ai_client = AnthropicClient.init(allocator, api_key);
defer ai_client.deinit();

var response = try ai_client.sendMessage(
    "claude-3-sonnet-20240229",
    "Analyze this trading algorithm...",
    1000,
);
defer response.deinit();

// Use response.content for your application logic

This example proves zig-http-sentinel's versatility across industries—from algorithmic trading to AI applications, delivering consistent enterprise-grade performance.

Example: Enterprise NATS JetStream Bridge

Demonstrating production messaging infrastructure integration with NATS JetStream via HTTP gateway - proving enterprise messaging capabilities.

Features

• V-Omega Protocol Compliance: Full integration with canonical V-Omega message patterns
• JetStream Management: Stream and consumer lifecycle operations
• High-Performance Publishing: Enterprise-grade message publishing with acknowledgments
• Telemetry Integration: Compatible with existing NATS infrastructure (172.191.60.219:4222)
• Multi-Domain Support: AI, HPC, and Quantum domain message routing

Run the NATS Demo

zig build nats-demo

V-Omega Message Pattern

// Canonical V-Omega message structure
vomega.{theater}.{domain}.{application}.{action}

// Examples:
vomega.azure.ai.hydra-chimera.telemetry.batch_complete
vomega.gcp.hpc.nuclear-fire-hose.telemetry.pps_report
vomega.aws.quantum.jetstream.telemetry.throughput

Integration Examples

const NatsJetStreamClient = @import("nats_bridge.zig").NatsJetStreamClient;

var client = NatsJetStreamClient.init(allocator, "172.191.60.219", 4222, "azure");
defer client.deinit();

// Publish AI telemetry
const ai_payload = std.json.Value{ .object = payload_map };
var response = try client.publishVOmegaMessage(
    "ai", "hydra-chimera", "telemetry.batch_complete", ai_payload
);
defer response.deinit();

// Create enterprise stream
const config = StreamConfig{ .max_msgs = 1000000, .storage = "file" };
var stream = try client.createVOmegaStream("quantum", "jetstream", config);
defer stream.deinit();

// Pull message batches
var batch = try client.pullMessages("VOMEGA_AZURE_AI", "processor", 100, 5000);
defer batch.deinit();

Supported Operations

• Stream Management: Create, configure, and monitor JetStream streams
• Consumer Operations: Durable consumers with acknowledgment policies
• Message Publishing: V-Omega compliant message publishing with sequence tracking
• Batch Processing: High-throughput message pulling and processing
• Monitoring: Stream statistics and consumer health metrics

This integration demonstrates zig-http-sentinel's capability to bridge HTTP and enterprise messaging systems, enabling hybrid architectures with consistent performance patterns.

Testing

Run the test suite:

zig build test

Requirements

• Zig 0.16.0 or later
• No external dependencies

Performance

This library has been optimized for high-throughput scenarios and has been tested under production loads handling thousands of requests per second.

Contributing

Contributions are welcome! Please ensure:

1. All tests pass
2. Code follows Zig style conventions
3. New features include tests
4. Documentation is updated

License

MIT License - See LICENSE file for details

Copyright © 2025 QUANTUM ENCODING LTD
Website: https://quantumencoding.io
Contact: [email protected]

Acknowledgments

This library emerged from real-world production needs and represents lessons learned from building high-performance trading systems at QUANTUM ENCODING LTD.