Copilot desktop app activity stopped being attributed to `copilot_app` in usage metrics
还没有人认领这个 Issue。
评估
- 难度
- 4/5
- 预计耗时
- 3-5 天
- 新手友好度
- 35/100
- Issue 类型
- 缺陷
- 描述清晰度
- 描述清楚
- 活跃度
- 活跃
- 技术栈
- github
- 领域
- analytics
调研方向
No source file or test is identified in the issue. Start by reproducing the organization users-1-day report for a desktop app user and compare the app and CLI attribution fields; the issue is done when desktop activity is attributed to copilot_app, retaining CLI attribution only when both surfaces are involved.
由索引模型根据 Issue 内容生成。
描述
Short summary
After a Copilot desktop app update, organization usage reports stopped attributing activity from some desktop app users to the copilot_app surface.
Affected version or release
1.1.21
Installation context
All people in the org using Github Copilot app.
What happened?
After a Copilot desktop app update, organization usage reports stopped attributing activity from some desktop app users to the copilot_app surface.
Before the update, desktop app sessions produced all expected signals:
used_copilot_app: truetotals_by_copilot_app- A
copilot_appentry intotals_by_feature - In some cases, simultaneous CLI attribution because the app uses the CLI runtime internally
After the update, the same users’ activity continued to be reported, but only as CLI activity:
used_cli: truetotals_by_clipresentcopilot_clifeature activity presentused_copilot_app: falsetotals_by_copilot_appabsent- No
copilot_appfeature entry
❗ Comparing weekdays before and after the transition:
- Reported Copilot App active users fell 78%
- Copilot App requests fell 94%
- Copilot App generations fell 96%
- CLI active users increased 42%
- CLI requests increased 97%
- CLI prompts increased 215%
- CLI-only attribution increased 196%
- Simultaneous app-and-CLI attribution fell 89%
The organization-level report reconciles exactly with the per-user rows, so this is not a downstream dashboard or aggregation problem. The incorrect attribution originates in GitHub’s telemetry or reporting pipeline.
Steps to reproduce
- Use the GitHub Copilot desktop app through an organization-managed Copilot license.
- Download the organization
users-1-dayusage report. - Locate the user’s activity for that day.
- Observe that desktop app activity is reported under
used_cliandtotals_by_cli. - Observe that
used_copilot_appisfalse, whiletotals_by_copilot_appand thecopilot_appfeature row are absent.
Expected behavior
Activity initiated in the GitHub Copilot desktop app must be attributed to the copilot_app surface, even if the app uses the Copilot CLI runtime internally.
If both surfaces are intentionally involved, the report should retain both signals, as it did previously and as GitHub’s documented example schema supports.
Additional context
‼️ This is not acceptable for a paid organization product. We rely on these metrics to measure Copilot App adoption and justify substantial licensing and AI-credit expenditure. Reporting desktop app usage as CLI-only materially misrepresents product adoption and makes the organization metrics unreliable.
- 主要语言
- 没有语言数据
- 星标
- 2.1k
- 派生
- 157
- PR 合并指标
- 30 天内没有已合并 PR
贡献指南
从这里开始
- 先读完整个 Issue,再读项目的贡献指南。
- 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
- Fork 仓库,在一个分支上完成修改。
- 提交 Pull Request,并在描述里引用这个 Issue 编号。
github/app 的其他 Issue
-
triage
难度 2/5 1-3 小时 新手友好度 70/100
-
难度 2/5 1-3 小时 新手友好度 65/100
-
难度 2/5 1-3 小时 新手友好度 70/100
-
难度 2/5 1-3 小时 新手友好度 70/100
-
难度 2/5 1-3 小时 新手友好度 70/100
相似的 Issue
-
难度 2/5 1-3 小时 新手友好度 91/100
-
feature
难度 2/5 1-3 小时 新手友好度 84/100
-
bug :bug:
难度 2/5 1-3 小时 新手友好度 82/100
-
难度 2/5 1-3 小时 新手友好度 85/100
epiforecasts/BVDOutbreakSize#888 · 1 条评论 ·
-
enhancement P3: gap query
难度 2/5 1-3 小时 新手友好度 75/100
jeffknupp/association#208 ·