Implement `output.cost_tracking` (per-GPU rates, alert / halt thresholds, audit and webhook events) — or retire the promise
维护者通常 1 天内回复
还没有人认领这个 Issue。
评估
调研方向
Start with the deferral row at docs/roadmap/risks-and-decisions.md:94 and the GPU-cost manual pages (EN and TR) that advertise output.cost_tracking. Confirm that ForgeConfig in forgelm/ has only the scalar training.gpu_cost_per_hour field. The maintainers have not chosen a path, so first ask whether to build the block or drop the promise. Done means either the schema, audit event, webhook event and EN/TR docs exist, or the manual and roadmap drop the promise. In both cases the row leaves the deferral table and the CHANGELOG names F-PR29-A2-04.
由索引模型根据 Issue 内容生成。
描述
Deferred work ·
F-PR29-A2-04· severity Medium · cost ~1 day · recorded indocs/roadmap/risks-and-decisions.md:94
Summary
The GPU-cost manual (EN + TR) advertises an output.cost_tracking block with per-GPU rates, a currency and alert / halt thresholds, but ForgeConfig has no such field; only the scalar training.gpu_cost_per_hour exists. The PR #29 review deferred it because it needs a CostTrackingConfig model, an audit event when a threshold is crossed and a webhook lifecycle event (which extends the published webhook vocabulary). The documentation issues only correct the pages, so the build-or-drop decision is still open.
Condition or budget
None recorded (the row still names v0.6.x).
State today (re-checked 2026-10-09)
No cost_tracking or CostTrackingConfig exists in forgelm/; the manual labels its examples as forward-looking placeholders.
Related
#428, #241, #429 and #444 correct the cost documentation; none asks for the implementation.
Acceptance criteria
- Either the block exists (schema, audit event, webhook event, EN/TR docs) or the manual and the roadmap drop the promise and the row is closed with that reason.
- The row is removed from the deferral table and the CHANGELOG names its ID (the table's removal contract).
Recorded in docs/roadmap/risks-and-decisions.md (read at f94595f) and re-checked against the code on 2026-10-09. The row leaves that table only when the fix is on main, the CHANGELOG names its ID and any promised guard exists.
- 主要语言
- Python
- 星标
- 9
- 派生
- 1
- 平均合并
- 8 小时 47 分钟
- 30 天内合并 PR
- 47
环境准备
- 提供 Dockerfile 或 Docker Compose 文件
- 有 Pull Request 模板
- 阅读贡献指南
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。
HodeTech/ForgeLM 的其他 Issue
-
area: site bug good first issue severity: low wave: 4
难度 2/5 1-3 小时 新手友好度 75/100
HodeTech/ForgeLM#961 · 1 条评论 ·
维护者通常 1 天内回复
-
area: dev-tooling bug severity: low source: roadmap wave: 4
难度 2/5 1-3 小时 新手友好度 70/100
维护者通常 1 天内回复
-
Trend-tracking jq recipe prints evaluation_completed=false as 'n/a', because jq's // operator treats false like null (F-W1-DOCS-TRAIN-16)可能已有人在做 @yuefdev 于 4 天前认领。 未关闭documentation good first issue severity: low source: review-2026-09 wave: 4
难度 1/5 1-3 小时 新手友好度 80/100
维护者通常 1 天内回复
-
documentation severity: medium source: review-2026-09 wave: 3
难度 2/5 1-3 小时 新手友好度 72/100
维护者通常 1 天内回复
-
documentation good first issue severity: medium source: review-2026-09 wave: 3
难度 2/5 1-3 小时 新手友好度 68/100
维护者通常 1 天内回复
相似的 Issue
-
Validate non-divisible effective batch sizes in the Unsloth LoRA backend可能已有人在做 @debesh-biswas 今天认领。 未关闭
难度 2/5 1-3 小时 新手友好度 70/100
-
难度 1/5 1 小时以内 新手友好度 88/100
reflex-dev/reflex#7634 · 1 条评论 ·
维护者通常 2 天内回复
-
documentation
难度 1/5 1-3 小时 新手友好度 82/100
topoteretes/cognee#6239 ·
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 78/100
ranaroussi/yfinance#3009 ·
维护者通常 1 天内回复
-
Test discovery aborts on GPU builds without cuStateVec: check_cuStateVec expects a failed result but set_config() now raises RuntimeError可能已有人在做 @dmeliksetian 今天认领。 未关闭bug
难度 2/5 1-3 小时 新手友好度 84/100
Qiskit/qiskit-aer#2495 ·
维护者通常 1 天内回复