tsenart/vegeta
Vegeta
Vegeta is an HTTP load testing tool and library designed to measure the performance and stability of web services. It functions as a command-line utility, a programmable package for integration into other applications, and a distributed load generator capable of splitting request rates across multiple machines.
The tool is distinguished by its constant-rate request scheduler, which dispatches requests at a fixed frequency regardless of target response times. It employs lazy target streaming to maintain low memory usage during large tests and uses a binary-encoded storage format to minimize disk I/O during high-throughput execution.
Beyond request generation, it provides a performance analysis toolkit for evaluating request latency, success rates, and throughput. This includes the ability to generate text histograms, JSON reports, and interactive HTML time-series plots. The system also integrates with Prometheus and Grafana by exposing real-time metrics via an HTTP endpoint.
Target definitions can be managed through structured schemas, plain text files, or dynamic streaming inputs.
Features
- HTTP Load Testing - Provides a high-volume HTTP request generator to measure the performance and stability of web services.
- Libraries - Provides a programmable package for integrating high-performance HTTP request generation into other applications.
- Command-Line HTTP Clients - Ships a command-line utility for sending a constant rate of HTTP requests to measure latency.
- Request Rate Controllers - Implements a constant-rate request scheduler to maintain a fixed frequency of requests regardless of target response times.
- Constant-Rate Request Schedulers - Implements a timing loop to dispatch HTTP requests at a fixed frequency regardless of the target response time.
- Distributed Load Generation - Scales load tests across multiple machines to simulate high-scale traffic and bypass single-node hardware limitations.
- Performance Analysis - Evaluates request latency, success rates, and throughput using reports and histograms to identify bottlenecks.
- Performance Testing - Subjects web services to high-volume traffic to evaluate stability and responsiveness under pressure.
- Load Testing Tools - Sends a configurable rate of requests to target URLs over a specified duration to measure performance.
- Steady-State Generation - Generates a steady stream of HTTP requests at a fixed frequency to measure system performance.
- Results Analysis - Generates text, JSON, and histogram reports from captured data to evaluate latency and success rates.
- Binary Serialization Formats - Writes raw test metrics to a compact binary format to minimize disk I/O and storage space.
- Time-Series Visualizers - Creates interactive HTML plots and time series graphs to analyze response time changes over the duration of a test.
- Lazy Target Streaming - Reads request definitions from files or standard input as a stream to keep memory usage low during large tests.
- Latency - Calculates success ratios, throughput, and latency percentiles from result files for performance evaluation.
- Metric Aggregation & Downsampling - Reduces large result sets into smaller data points using moving averages to render interactive plots.
- File-Based Target Loading - Reads target lists from files or standard input and distributes requests across them using a round-robin strategy.
- Round-Robin Target Distribution - Cycles through a list of target endpoints sequentially to distribute load evenly across multiple specified URLs.
- Structured Target Specifications - Allows defining request methods, URLs, bodies, and headers via a structured schema for execution.
- Request Payloads - Supports defining target requests using JSON, HTTP, or plain-text formats to specify payloads.
- Decoupled Architectures - Separates the request execution phase from the reporting phase using interchangeable result encoders and decoders.
- Performance Reporting - Produces HTML5 plots, text histograms, and CSV records to analyze request latency and throughput.
- Prometheus Exporters - Exposes internal performance counters via an HTTP endpoint for Prometheus scraping.
- Dynamic Target Streaming - Supports streaming request definitions from external programs or files in HTTP or JSON format to simulate varied traffic.
- Test Reporters - Transforms raw load test data into structured formats to simplify downstream reporting.
- Performance Test Visualizations - Transforms raw test data into visual plots for detailed analysis of latency and throughput over time.
- Performance Testing and Analysis - Provides a toolkit for converting load test results into CSV, JSON, and interactive HTML latency plots.
- Load Ramping - Executes tests at increasing request rates to analyze how latency and success rates change under escalating stress.
- Test Result Converters - Transforming result files between various data encodings to facilitate external analysis and data processing.
- Web Traffic Simulators - Mimics high-concurrency user behavior to test the resilience and stability of web services.
- Target Endpoint Mapping - Specifies target URLs and methods using plain text, JSON, or streaming input to determine endpoints for stress testing.
- Infrastructure Tools - HTTP load testing tool and library.
- Command Line Tools - Listed in the “Command Line Tools” section of the The Book Of Secret Knowledge awesome list.
- Load Testing Tools - Versatile HTTP load testing library and CLI tool.
- HTTP Security - Load testing tool for constant throughput benchmarking.
