robbie-microsoft/Continuous Benchmark
Continuous Benchmark using GitHub pages as dash board for keeping performance
View on GitHubTrust Signals
- Scorecard Score
- not yet scored
- Maintenance Recency
- Stale
- License
- None
- Runtime
- Deprecated runtime
Inputs
| name | description | required | default |
|---|---|---|---|
| name | Name of the benchmark. This value must be identical among all benchmarks | yes | Benchmark |
| tool | Tool to use get benchmark output. One of "cargo", "go", "benchmarkjs", "pytest", "googlecpp", "catch2", "julia", "benchmarkdotnet", "customBiggerIsBetter", "customSmallerIsBetter" | yes | — |
| output-file-path | A path to file which contains the benchmark output | yes | — |
| gh-pages-branch | Branch for gh-pages | yes | gh-pages |
| gh-repository | Url to an optional different repository to store benchmark results | no | — |
| benchmark-data-dir-path | Path to directory which contains benchmark files on GitHub pages branch | yes | dev/bench |
| github-token | GitHub API token to pull/push GitHub pages branch and deploy GitHub pages. For public repository, this must be personal access token for now. Please read README.md for more details | no | — |
| ref | optional Ref to use when finding commit | no | — |
| auto-push | Push GitHub Pages branch to remote automatically. This option requires github-token input | no | false |
| skip-fetch-gh-pages | Skip pulling GitHub Pages branch before generating an auto commit | no | false |
| comment-always | Leave a comment with benchmark result comparison. To enable this feature, github-token input must be given as well | no | false |
| summary-always | Leave a job summary with benchmark result comparison | no | false |
| save-data-file | Save the benchmark data to external file | no | true |
| comment-on-alert | Leave an alert comment when current benchmark result is worse than previous. Threshold is specified with alert-comment-threshold input. To enable this feature, github-token input must be given as well | no | false |
| alert-threshold | Threshold which determines if an alert should happen or not. Percentage value such as "150%". For example, 150% means that an alert happens when current benchmark result is 1.5x worse than previous | no | 200% |
| fail-on-alert | Workflow fails when alert comment happens | no | false |
| fail-threshold | Threshold which determines if the current workflow fails. Format is the same as alert-threshold input. If this value is not specified, the same value as alert-threshold is used | no | — |
| alert-comment-cc-users | Comma separated GitHub user names which start with @ (e.g. "@foo,@bar"). They will be mentioned in commit comment for alert. | no | — |
| external-data-json-path | JSON data file for storing benchmark results. When this input is set, github-action-benchmark no longer uses Git branch to store data. Instead, it reads and appends benchmark data from/to the file. User must store the file anywhere | no | — |
| max-items-in-chart | Max data points in a benchmark chart to avoid making the chart too busy. Value must be unsigned integer. No limit by default | no | — |
Outputs
no outputs