throttlecrab-server

A high-performance rate limiting server with multiple protocol support, built on throttlecrab.

Features

• Multiple protocols: Native binary, HTTP (JSON), and gRPC
• High performance: Lock-free shared state with Tokio async runtime
• Production ready: Health checks, configurable logging, systemd support
• Flexible deployment: Docker, binary, or source installation
• Shared rate limiter: All protocols share the same store for consistent limits

Installation

Install the server binary with cargo:

cargo install throttlecrab-server

Or build from source:

git clone https://github.com/lazureykis/throttlecrab
cd throttlecrab/throttlecrab-server
cargo build --release
./target/release/throttlecrab-server

Usage

Start the throttlecrab server (at least one transport must be specified):

# Run with HTTP transport
throttlecrab-server --http

# Run with HTTP transport on custom port
throttlecrab-server --http --http-port 7070

# Run multiple transports simultaneously
throttlecrab-server --http --grpc

# Specify different hosts and ports for each transport
throttlecrab-server --http --http-host 0.0.0.0 --http-port 8080 \
                    --grpc --grpc-port 50051

# Configure store type and parameters
throttlecrab-server --http --store adaptive \
                    --store-min-interval 5 \
                    --store-max-interval 300 \
                    --store-max-operations 1000000

# Use periodic store with custom cleanup interval
throttlecrab-server --http --store periodic --store-cleanup-interval 600

# Use probabilistic store
throttlecrab-server --grpc --store probabilistic --store-cleanup-probability 5000

# Set custom buffer size and log level
throttlecrab-server --http --buffer-size 50000 --log-level debug

Environment Variables

All CLI arguments can be configured via environment variables with the THROTTLECRAB_ prefix:

# Transport configuration
export THROTTLECRAB_HTTP=true
export THROTTLECRAB_HTTP_HOST=0.0.0.0
export THROTTLECRAB_HTTP_PORT=8080

# Store configuration
export THROTTLECRAB_STORE=adaptive
export THROTTLECRAB_STORE_CAPACITY=200000
export THROTTLECRAB_STORE_MIN_INTERVAL=10

# General configuration
export THROTTLECRAB_BUFFER_SIZE=100000
export THROTTLECRAB_LOG_LEVEL=info

# CLI arguments override environment variables
THROTTLECRAB_HTTP_PORT=8080 throttlecrab-server --http --http-port 7070
# Server will use port 7070 (CLI takes precedence)

Transport Comparison

Protocol Documentation

Native Protocol (Recommended)

Fixed-size binary protocol with minimal overhead:

• Request: 88 bytes (including up to 64-byte key)
• Response: 40 bytes
• No serialization overhead
• Direct memory layout

For best performance, implement the native protocol in your application or use established HTTP clients with connection pooling.

HTTP REST API

Endpoint: POST /throttle

Request Body (JSON):

{
  "key": "user:123",
  "max_burst": 10,
  "count_per_period": 100,
  "period": 60,
  "quantity": 1,
  "timestamp": 1234567890123456789
}

Note: timestamp is optional (Unix nanoseconds). If not provided, the server uses the current time.

Response (JSON):

{
  "allowed": true,
  "limit": 10,
  "remaining": 9,
  "reset_after": 60,
  "retry_after": 0
}

gRPC Protocol

See proto/throttlecrab.proto for the service definition. Use any gRPC client library to connect.

Client Integration

Rust

Use established HTTP clients like reqwest for reliable production deployments:

[dependencies]
reqwest = { version = "0.12", features = ["json"] }

Other Languages

• Go: Use gRPC with generated client
• Python: Use HTTP/JSON API
• Node.js: Use HTTP/JSON API
• Java: Use gRPC with generated client

See examples/ directory for implementation examples.

Running Benchmarks

Criterion Benchmarks

The project includes several Criterion benchmarks that require running servers. Use the provided script:

# Run all benchmarks
./run-criterion-benchmarks.sh

# Run specific benchmark
./run-criterion-benchmarks.sh tcp_throughput
./run-criterion-benchmarks.sh connection_pool
./run-criterion-benchmarks.sh protocol_comparison
./run-criterion-benchmarks.sh grpc_throughput

The script will:

1. Build the server in release mode
2. Start required servers (native on port 9092, gRPC on port 9093)
3. Run the benchmarks
4. Clean up servers on exit

Results are saved in target/criterion/.

Production Deployment

Performance Tuning

# Optimal settings for production
throttlecrab-server \
    --native --native-port 9090 \
    --store adaptive \
    --store-capacity 1000000 \
    --buffer-size 100000 \
    --log-level warn

Monitoring

• Health endpoint: GET /health (available on HTTP port)
• Logs: Structured logging with configurable levels
• Metrics: Performance metrics in debug/trace logs

Store Configuration

Example Configurations

High-throughput API

throttlecrab-server --native \
    --store adaptive \
    --store-capacity 5000000 \
    --buffer-size 500000

Web Service

throttlecrab-server --http --http-port 8080 \
    --store periodic \
    --store-cleanup-interval 300

Microservices

throttlecrab-server --grpc --native \
    --grpc-port 50051 \
    --native-port 9090 \
    --store probabilistic

License

Licensed under either of:

• Apache License, Version 2.0 (LICENSE-APACHE or http://www.apache.org/licenses/LICENSE-2.0)
• MIT license (LICENSE-MIT or http://opensource.org/licenses/MIT)

at your option.

Contribution

Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in the work by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.