Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/kitlangton/jmh-benchmark-action
A GitHub CI Action for automating the comparison of JMH benchmarking results.
https://github.com/kitlangton/jmh-benchmark-action
Last synced: 2 months ago
JSON representation
A GitHub CI Action for automating the comparison of JMH benchmarking results.
- Host: GitHub
- URL: https://github.com/kitlangton/jmh-benchmark-action
- Owner: kitlangton
- Created: 2023-08-04T11:23:43.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2023-10-19T17:26:09.000Z (about 1 year ago)
- Last Synced: 2024-10-11T08:45:22.460Z (3 months ago)
- Language: Scala
- Size: 28.2 MB
- Stars: 4
- Watchers: 2
- Forks: 1
- Open Issues: 5
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# JMH Benchmarking Action
This GitHub action is designed for running and comparing JMH benchmarks. It runs your benchmarks and checks for any
performance regressions beyond a configurable threshold. When it detects a regression, it can either fail or provide a
warning based on the `failOnRegression` configuration.If the GitHub event is a pull request, it adds a comment to the PR showing the benchmark comparison. If it's a push
event, it updates the saved benchmark data with the new data (stored in a separate `internal/benchmark-data` branch).### Example Automated Pull Request Comment
![CleanShot 2023-08-04 at 12 08 55@2x](https://github.com/kitlangton/jmh-benchmark-action/assets/7587245/1e06415a-2966-4c06-b4f5-9d40bea0b271)
Taken from this [Example Project](https://github.com/kitlangton/jmh-benchmark-action-example-project).
## Usage
To use the JMH Benchmarking Action, you need to include it in your GitHub workflow. An example workflow configuration
can be found below:```yaml
name: Benchmark
on:
push:
branches:
- main
pull_request:
branches:
- mainpermissions:
contents: write
pull-requests: writejobs:
build:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v2
- name: Set up JDK 11
uses: actions/setup-java@v2
with:
java-version: '11'
distribution: 'adopt'
- name: Cache SBT dependencies
uses: coursier/cache-action@v6
- name: Build and run JMH benchmark
run: |
sbt clean compile
sbt 'benchmarks/jmh:run -i 3 -wi 3 -f1 -t1 -rf json -rff output.json .*'
- name: JMH Benchmark Action
uses: kitlangton/jmh-benchmark-action@main
with:
jmh-output-path: benchmarks/output.json
github-token: ${{ secrets.GITHUB_TOKEN }}
```When you set up the action, you will need to provide the path to your benchmark output and your GitHub token.
## Configuration
You can configure the action by providing additional parameters. The available parameters are:
- `jmh-output-path`: The file path where your JMH benchmark results are written (e.g., `benchmarks/output.json`).
- `github-token`: Your GitHub token, (e.g., `${{ secrets.GITHUB_TOKEN }}`).
- `fail-on-regression`: (Optional) A boolean that indicates whether the action should fail when it detects regressions
beyond a certain threshold (defaults to false).
- `failure-threshold`: (Optional) The change beyond which a performance decrease is considered a regression (defaults to
-0.05).