Welcome to the edb-benchmarks repository!

This repository contains workflows for benchmarks that have been previously developed and are ready for execution.

Getting Started

Benchmarks can be executed through a dedicated Python virtual environment.

Components needed by host:

• terraform >= 1.3.6
• edb-terraform >= 1.5.1
• AWSCli
• ansible-core
• edb-ansible
• tpaexec

Benchmark folder structure:

Each benchmark should follow the directory structure listed below with a numeric prefix and a step name:

• 00_validate - Validates the parameters selected for the benchmark execution. If the validation fails, the entire benchmark fails.
• 01_provision - Provisions the infrastructure required for executing the benchmark
• 02_deploy - Sets up and configures the provisioned infrastructure
• 03_prepare - Prepares and sets up the benchmark execution
• 04_execute - Executes the benchmark
• 05_push_results - Uploads the results from the benchmark into AWS S3 Bucket
• 06_unprovision - Destroys the previously provisioned infrastructure

Each directory contains a run.sh script file that executes the ansible playbooks required for the benchmark

Executing a benchmark by hand

Clone the repo to a local machine.

$ git clone https://github.com/EnterpriseDB/edb-benchmarks.git

From the terminal, set environment variables required by the cloud service provider.

If desired, make changes to the type or size of the database by editing the environment file: environment.sh.

Source the environment file.

source <benchmark-directory>/environment.sh

Execute <benchmark-directory>/00_validate/run.sh

Execute <benchmark-directory>/01_provision/run.sh

Wait for the 'Apply Complete' success message.

Execute <benchmark-directory>/02_deploy/run.sh

If desired, export additional environment variables manually.

Execute <benchmark-directory>/03_prepare/run.sh

Execute <benchmark-directory>/04_provision/run.sh

When ready to terminate the instance, execute <benchmark-directory>/06_unprovision/run.sh

Contributing:

We welcome contributions! If you have benchmarks that you would like to share follow the steps below:

• Simply fork this repository
• Add your benchmarks
• Submit a pull request.

Please ensure that the code is well-documented and follows our Guidelines below.

Guidelines:

When submitting benchmarks, please follow these guidelines:

• Use clear and concise names for the benchmark files and directories
• Include a README.md file in the main directory. The file content should explain:
  • What the benchmark does
  • How to run it
  • Relevant details related to the benchmark
  • Make sure that the benchmarks are reproducible

License

The license is available here