parkbench

Parker Benchmark Tool

A benchmarking tool for testing the performance of Parker services, available in both Python and Go implementations. The tool supports both HTTP and gRPC protocols and provides detailed performance metrics.

Prerequisites

• Go 1.21 or later (for Go benchmarks)
• Python 3.8+ (for Python benchmarks)
• Make

Installation

1. Clone the repository
2. For Python, install dependencies and generate protobuf files:

make setup
# The `make setup` command will:
# - Create a Python virtual environment
# - Install required Python packages
# - Install gRPC tools

Configuration

Create a config.json file with the following structure:

{
    "grpcAddress": "aws-us-west-1-001.api.parkerdb.com:50051",
    "httpAddress": "https://aws-us-west-1-001.api.parkerdb.com",
    "csv": "test_data.csv",
    "account": "your_account",
    "table": "your_table",
    "jwt": "your_jwt_token",
    "concurrency": 10,
    "repeat": 1
}

Configuration Options

• csv: Path to the CSV file containing IDs to query
• account: Your Parker account name
• table: The table to query
• httpAddress: HTTP server address (for HTTP protocol)
• grpcAddress: gRPC server address (for gRPC protocol)
• jwt: JWT token for authentication (optional)
• concurrency: Number of concurrent workers
• repeat: Number of times to repeat the benchmark

Usage

The benchmark tool can be run using Make targets:

Python Implementation

# Run HTTP benchmark using Python
make benchmark-http

# Run gRPC benchmark using Python
make benchmark-grpc

Go Implementation

# Run HTTP benchmark using Go
make benchmark-go-http

# Run gRPC benchmark using Go
make benchmark-go-grpc

Output

The tool will display:

• Total number of requests processed
• Average latency
• P50, P95, and P99 latency percentiles
• Requests per second

Example Output

Running benchmark: 100%|██████████| 10/10 [00:05<00:00, 1.98it/s]

Benchmark Results:
Total Requests: 1000
Average Latency: 50.23ms
P50 Latency: 45.67ms
P95 Latency: 75.89ms
P99 Latency: 89.12ms
Requests per Second: 198.45

Development

Building and Generating Code

# Generate Python protobuf files (included in make setup)
make generate-python

# Generate Go protobuf files
make generate-go

# Clean generated files and virtual environment
make clean

The make clean command will remove:

• Generated Python protobuf files
• Python __pycache__ directories
• Generated Go protobuf files
• Python virtual environment

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.