Quisby is a tool to provide first view into the results from various benchmarks such as linpack, streams, fio etc. It doesn't aim to replace existing data viz tool but rather to provide a simplified view to the data with basic metric to understand the benchmark results from a higher level view. For detailed view, there are other tools such as pbench-dashboard, js-charts etc at hand.
Bechmarks currently supported:
| Benchmark | Source data |
|---|---|
| linpack | Benchmark result |
| streams | Summary result |
| uperf | Summary csv result |
| specjbb | Benchmark result |
| pig | Benchmark result |
| hammerDB | Benchmark result |
| fio | pbench result |
| autohpl | Summary result |
| aim | Benchmark result |
| etcd | pbench result |
| reboot | Benchmark result |
| speccpu | Benchmark result |
It extracts data from benchmark results file or summary results produced by wrapper benchmark programs and move that results to Google Sheet via sheets API V4.
Run
$ pip install quisbyIt takes in an input file with list of location to the test results. The location file will look like:
test: results_linpack
</path/to/results>
...
test: pbench_fio
<http url with results>
...
Then you can run
$ quisby process --os-type <add-here> --os-release <add-here> --cloud-type <add-here> location_file`For more information on options, run:
$ quisby -hThat's it. It will return a google sheet. Visit the google sheet page and you will see a newly created spreadsheet with the data populated and graphed.
If you want to compare two different OS release of similar system type then there are scripts that will help you to create a spreadsheet for the same.
and then run:
quisby compare --test-name <benchmark-name-(optional)> --spreadsheets <spreadsheet1,spreadsheet2>and it would return a newly created spreadsheet with the comparison data.
#Clone the repo
git clone [email protected]:sourabhtk37/data-to-sheet.git
# Installation
source ./install.sh
(optional, for configuring aws and/or azure cli)
source ./install -aws -azurequisby is now configured for local development, to run it:
$ poetry run quisby [options] <location_file>Create issues and create a seperate feature branch to work on it. Push the changes to your clone repo and then create a pull request to the master branch of the origin repo.