GitHub Copilot’s usage‑metrics API now includes optional fields that record activity in the dedicated VS Code Agents window, both in aggregate enterprise/organization reports and in per‑user reports. Engineers can now query daily active user counts, session totals, and message totals for the Agents UI, giving a concrete signal of adoption and engagement across teams.
New aggregate and per‑user fields
The API adds two optional elements to the 1‑day and 28‑day report windows:
- daily_active_vscode_agent_users: a count of distinct users who interacted with the VS Code Agents window on a given day.
- totals_by_vscode_agent: an object that aggregates
session_countandtotal_user_messagesfor the window.
For individual‑user reports the same period granularity now returns:
- used_vscode_agent: a boolean flag indicating whether the user opened the Agents window.
- totals_by_vscode_agent: per‑user
session_countandtotal_user_messagesvalues.
Why the data matters to engineers
These metrics isolate the dedicated Agents UI from the broader “Agent Mode” that runs inside the editor window. By separating the two, teams can track the specific adoption path of the standalone Agents experience, compare it against other Copilot usage, and spot trends such as spikes in session volume that may correlate with new feature roll‑outs or training initiatives.
Operational and security considerations
Access to the new fields respects the existing permission model: only enterprise owners, billing managers, organization owners, or custom roles with the View Copilot Metrics privilege can retrieve them, and the Copilot usage‑metrics policy must be enabled. Because the fields are optional, missing data will appear as absent or null, preserving backward compatibility with existing dashboards. Practitioners should verify that their metric collection pipelines can handle nullable values without breaking downstream alerts or visualisations.
Related CloudNinjas coverage: AI engineering.
What This Means For Practitioners
Start by extending any internal Copilot observability dashboards to ingest daily_active_vscode_agent_users and the per‑user totals_by_vscode_agent payloads. Use the daily active count as a leading indicator of Agents adoption, and correlate session and message totals with capacity planning for backend services that support the Agents feature. Ensure that role assignments and the usage‑metrics policy are reviewed so that the right stakeholders have visibility without over‑exposing data. Finally, monitor the optional fields for null values; a sudden increase in nulls could indicate a client‑side upgrade lag or a configuration drift that needs remediation.
