tracerbench/@tracerbench/tracerbench-compare-action

Perform commit-to-commit performance analysis using tracerbench

View on GitHub

Trust Signals

Scorecard Score
not yet scored
Maintenance Recency
Stale
License
None
namedescriptionrequireddefault
use-yarnWhether to install dependencies using yarnno
use-pnpmWhether to install dependencies using pnpmno
pkg-managerpackage manager to use, defaults to npm unless use-yarn or use-pnpm are set which take precedenceno
repo-tokenGithub Auth Token with permissions for reading, creating comments on PRs as well as updating comments previously created by the actionno
control-shaSHA of the control to use for the comparison. Also used to checkout the correct files before build if build-control is true.
experiment-shaSHA of the experiment to use for the comparison. Also used to checkout the correct files before build if build-experiment is true.
build-controlWhether to run control-build-command to produce control-distno
build-experimentWhether to run experiment-build-command to produce experiment-distno
control-build-commandcommand to execute to produce control-dist if build-control is trueno
control-serve-commandember s --path=dist-control
experiment-build-commandcommand to execute to produce experiment-dist if build-experiment is trueno
experiment-serve-commandember s --path=dist-experiment
control-distthe location of the built dist for the control-sha (e.g. the output of ember build)no
experiment-distthe location of the built dist for the experiment-sha (e.g. the output of ember build)no
markersmarkers, comma separatedno
control-urlthe url to sample for the controlno
experiment-urlthe url to sample for the experimentno
regression-thresholdthreshold for estimated delta to consider failed (in addition to statistically significant)no
fidelitylow, high or number of samplesno
debugWhether to run tracerbench in debug modeno
is-ci-envWhether to run tracerbench in CI modeno
upload-tracesWhether to upload all trace data as artifacts after the runno
upload-resultsWhether to upload all generated reports as artifacts after the runno
clean-after-analyzewhether to cleanup after analysisno
scenariosjson list of scenarios keyed by name with their control+experiment urls and markersno
sample-timeoutnumber of seconds to wait for a sample. Default value is 30sno
browser-argsadditional browser args to pass to tracerbench, comma separatedno

no outputs