Cost analytics
Per-run attribution of GitHub Actions minutes and spend — by repository, workflow, runner type, and storage, with wasted-spend and forecasting.
/cost gives you the bill GitHub does not show: which repositories, pipelines, and job types actually consume the money.
What is on the page
- Daily & monthly breakdown — bar and area charts of spending trajectory over the selected window.
- Runner architecture breakdown — Linux (Ubuntu), Windows, macOS, and self-hosted usage with their per-minute multipliers.
- Wasted spend — minutes burned on jobs that failed or were cancelled: the cheapest savings available.
- Storage costs — artifact storage and cache usage against GitHub billing tiers.
- Attribution tables — sortable rankings of the most expensive repositories, workflows, and recurring cron jobs.
- Forecast — projected month-end bill based on the current burn rate.
- CSV export — billing-ready line items for finance or chargeback.
How costs are calculated
Billable minutes follow GitHub's published rates: a job's minutes are multiplied by the runner OS factor (Windows and macOS cost more than Linux). A run whose runner type cannot be determined contributes 0 to cost — Gitera never guesses a rate.
Numbers are estimates for planning and attribution. They can differ slightly from the GitHub invoice — see cost mismatch troubleshooting.
Typical workflow
- Sort the attribution table by cost → find the top spenders.
- Check the wasted-spend block → fix flaky pipelines with auto-retry or failure triage.
- Check storage → prune artifacts/caches approaching paid tiers.