GitHub Actions minutes overage versus self-hosted runner cost for low-traffic private repositories

The goal is to compare GitHub-hosted Actions minutes overage against self-hosted runner compute cost for a low-traffic private repository on GitHub.com.

GitHub-hosted runners consume included Actions minutes and bill overage per minute by operating system. Self-hosted runners avoid per-minute charges but typically require continuously provisioned compute unless automated start/stop is added, creating idle fixed cost. The break-even point depends on job frequency, duration, runner OS, existing idle capacity, and current plan allotment.

Pricing and included minutes change by plan and region, and self-hosted usage may still incur storage and other applicable charges. Current billing rates should be confirmed before comparing options.

What usage history and billing data are needed to calculate the break-even point? Which runner operating systems and job durations make self-hosted compute cost-effective? What method accounts for idle runner capacity when jobs run infrequently?