Cost & Billing

Cost & Billing separates provider configuration from actual consumption. It is read-only toward GitHub and preserves unavailable or unauthorized data as a capability state rather than displaying a false zero.

Overview and actual usage

Overview summarizes provider-reported gross and net metered cost, discounts, fixed-license lines, unclassified charges, unattributed usage, and a clearly labeled run-rate projection. Actual Usage can be filtered by date, organization, repository, product, and SKU and includes:

• GitHub Actions minutes and storage/cache usage when returned;
• Packages storage and bandwidth;
• Codespaces compute and storage;
• Copilot, AI-credit, and premium-request usage when GitHub returns it;
• Advanced Security and previously unknown provider products/SKUs; and
• gross, discount, and net quantities and amounts with daily freshness.

Repository attribution is shown only when GitHub supplies enough information. Organization-wide or otherwise unattributed charges remain visible and are not assigned to an arbitrary repository.

Budget settings and Copilot

Budget Settings shows configured provider controls: budget ID and type, product, every SKU, organization/repository/user scope, amount or unit, alert state and recipients, stop-at-limit behavior, freshness, and capability state. A repository detail view includes direct controls and applicable broader controls labeled as inherited; it does not repeat every workspace budget as a direct repository budget.

GitHub Copilot Business or Enterprise subscription plans, seat counts, assignment mode, feature policy, and seat-review signals are separate from metered AI-credit or premium-request budgets. “No configured metered Copilot or AI-credit budget returned by GitHub” means the budget endpoint returned no such configured control. It does not mean the organization has no Copilot plan or that usage is zero.

Coverage and optimization

Coverage identifies permission or enhanced-billing limitations, stale data, products without matching controls, repository attribution gaps, and Copilot subscription/seat-detail availability. Optimization findings are advisory and read-only: they can highlight seats for human review, recent cost increases, high-cost Actions repositories, and storage, cache, artifact, or bandwidth cleanup candidates.

CSV export uses the active filters. Budget comparisons always use the complete provider scope, and projections use the latest provider day rather than assuming missing days are zero.

History

Durable daily usage supports current-period trends. Comparing the current and at least two prior billing months in the interface is tracked in Azure DevOps item 8009 and is not yet presented as shipped functionality.