gbudjeakp/RunRight CI Resource Monitor

Monitor CI job resource usage and get right-sized machine recommendations

View on GitHub

Trust Signals

Scorecard Score
not yet scored
Maintenance Recency
Stale
License
None
namedescriptionrequireddefault
stepAction step to run: 'start' or 'stop' (standalone mode)no""
runCommand to wrap (wrapper mode — runs command with monitoring active)no""
durationMax monitoring duration (e.g. 30m). Defaults to unlimited.no0
intervalMetrics sampling interval (e.g. 5s)no5s
exportComma-separated export backends: file,otlp,prometheus,httpnofile
output-dirDirectory to write metrics filesno${{ github.workspace }}/.runright
job-idUnique job identifier (defaults to the GitHub job name)no${{ github.job }}
http-urlrunright backend URL for http exportno""
upload-artifactUpload metrics files as a GitHub Artifactnotrue
versionrunright binary version to downloadnolatest
github-tokenGitHub token used to post PR comments (defaults to the built-in token)no${{ github.token }}
pr-commentPost recommendations as a PR comment (true/false)notrue
providerFilter recommendations by cloud provider: aws, gcp, or github (default: all)no""
max-cost-per-hourBudget guard: fail the step if the detected machine's on-demand price exceeds this value (e.g. 0.50). Empty = disabled.no""
notify-slackSlack incoming webhook URL. When set, posts a per-job saving alert on completion.no""
notify-teamsMicrosoft Teams incoming webhook URL. When set, posts a per-job saving alert on completion.no""
dry-runObservation-only mode: collect metrics and log recommendations but skip PR comments and artifact uploads.nofalse
rollupCombine all RunRight PR comments for this workflow into a single comment table (one row per job).nofalse
namedescription
recommendation-jsonJSON array of machine recommendations
artifact-urlURL of the uploaded metrics artifact
detected-machineMachine type detected at runtime
detected-machine-confidenceConfidence score (0..1) for machine detection
detected-machine-confidence-levelConfidence label for machine detection: high, medium, low, unknown
detected-storage-typeStorage type from the detected catalog machine (for example ebs, ssd, pd-ssd)
runtime-storage-classBest-effort runtime storage class probe from the runner OS (ssd, hdd, unknown)
suggested-machineTop recommended machine ID