gurl

gurl = modern + curl + wrk

A modern HTTP benchmarking tool inspired by wrk, with the ability to parse curl commands for load testing.

Features

• 🚀 High-performance HTTP load testing
• 🔧 Parse curl commands with --parse-curl option
• 📊 Detailed statistics similar to wrk
• ⚡ Async I/O for maximum performance
• 🎯 Configurable connections, threads, and duration
• 📈 Latency distribution analysis

Installation

go install github.com/antlabs/gurl@latest

Or build from source:

git clone https://github.com/antlabs/gurl.git
cd gurl
go build -o gurl

Usage

Basic Usage

# Simple GET request
gurl -c 100 -d 30s http://example.com

# POST request with custom headers
gurl -c 50 -d 10s -X POST -H "Content-Type: application/json" -d '{"key":"value"}' http://api.example.com

Parse Curl Commands

# Parse a curl command and use it for benchmarking
gurl --parse-curl "curl -X POST -H 'Content-Type: application/json' -d '{\"name\":\"test\"}' http://api.example.com/users" -c 100 -d 30s

Options

• -c, --connections: Number of HTTP connections to keep open (default: 10)
• -d, --duration: Duration of test (default: 10s)
• -t, --threads: Number of threads to use (default: 2)
• -R, --rate: Work rate (requests/sec) 0=unlimited (default: 0)
• --timeout: Socket/request timeout (default: 30s)
• --parse-curl: Parse curl command and use it for benchmarking
• -X, --method: HTTP method (default: GET)
• -H, --header: HTTP header to add to request
• -d, --data: HTTP request body
• --content-type: Content-Type header
• -v, --verbose: Verbose output
• --latency: Print latency statistics

Examples

Basic Load Test

gurl -c 100 -d 30s -t 4 http://example.com

Output:

Running 30s test @ http://example.com
  4 threads and 100 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency    12.34ms   5.67ms  123.45ms    75.23%
    Req/Sec     2.34k     0.12k    3.45k     89.12%
  280000 requests in 30.00s, 45.67MB read
Requests/sec:   9333.33
Transfer/sec:     1.52MB

Load Test with Curl Command

gurl --parse-curl "curl -X POST -H 'Authorization: Bearer token123' -H 'Content-Type: application/json' -d '{\"user\":\"test\"}' https://api.example.com/login" -c 50 -d 10s --latency

Rate Limited Test

# Limit to 1000 requests per second
gurl -c 10 -d 60s -R 1000 http://example.com

Output Format

The output format is similar to wrk:

• Thread Stats: Average, standard deviation, and maximum latency
• Req/Sec: Requests per second statistics
• Total Summary: Total requests, duration, and data transferred
• Error Summary: Connection, read, write, and timeout errors (if any)
• Status Code Distribution: HTTP status code breakdown
• Latency Distribution: Percentile breakdown (with --latency flag)

Performance

gurl is designed for high performance:

• Async I/O using custom pulse library
• Efficient connection pooling
• Minimal memory allocation during testing
• Multi-threaded architecture

License

MIT License - see LICENSE file for details.