[Backlog Discovery] feat(backlog): automatic-workflow-rerun-for-transient-failures

[bestony]

[Backlog Discovery]

• Requirement title: 为工作流失败提供自动重试与自愈重跑
• Priority: P2
• Requirement file: backlog/20260221061845-automatic-workflow-rerun-for-transient-failures.md
• Dedupe key: automatic-workflow-rerun-for-transient-failures
• Source run: https://github.com/bestony/self/actions/runs/22251796835

---

[Backlog Discovery]

Update Record 2026-02-21 14:21:51 +08:00

Update summary:

• Added requirement to avoid reruns on deterministic failures alongside transient issues.

• Expanded core scenario to include global rerun frequency constraints.

• Refined solution to use failure classification rules, blacklists, and global rate limiting with backoff and caps.

• Updated acceptance criteria to require configurable failure classification/blacklists, global throttling, and a disable switch while keeping rerun logging.

• Status: committed

• Commit: 6e7fafa5f41bc711c4c79507be4b7fa8c5f53bd8

• Trigger: workflow_run.completed

• Comment: https://github.com/bestony/self/actions/runs/22251845722

• Actor: @github-actions[bot]

• Updated At (Asia/Shanghai): 2026-02-21 14:21:51 +08:00

---

[Backlog Discovery]

Update Record 2026-02-21 14:24:45 +08:00

Update summary:

• Added requirement to evaluate rerun success and false-positive rates during an observation period to refine failure classification and blacklist rules.

• Added placeholder for a 3‑month repository failure-type analysis to quantify impact and validate coverage of auto-rerun strategy.

• Expanded solution hypothesis to include a clear configuration entry point plus an observation/evaluation mechanism to calibrate rules.

• Extended delivery criteria with configurable observation period and metrics (success rate, false positives) to tune classification/blacklist.

• Status: committed

• Commit: 45027d541d5b8d4b76aaa2103bc9a39a430a6c03

• Trigger: workflow_run.completed

• Comment: https://github.com/bestony/self/actions/runs/22251883862

• Actor: @github-actions[bot]

• Updated At (Asia/Shanghai): 2026-02-21 14:24:45 +08:00

[gemini-code-assist]

Warning

You have reached your daily quota limit. Please wait up to 24 hours and I will start processing your requests again!

[github-actions]

[Reviewer Workflow]
Reviewer: Product Manager

需求价值评估

• 是否有价值: 有价值
• 优先级: P2
• Reviewer 视角结论: 对减少临时性失败中断和人工介入有直接价值，但需要补齐本仓库失败分布数据以确认收益边界。

价值点

• 提升自驱迭代稳定性：针对 runner/网络/外部服务短暂失败自动重跑，减少人工恢复等待。
• 降低运维成本：通过失败分类与黑名单避免无效重跑，减少重复排查与噪音。
• 风险控制可配置：全局限流、退避与开关机制能控制资源/速率风险并便于灰度验证。

风险与建议

• 证据不足：当前缺少近 3 个月失败类型与占比数据，建议补充统计以验证优先级与策略覆盖。
• 误判与循环风险：失败分类规则与黑名单可能滞后，建议先限定重跑次数与观察期指标并逐步放开。

[github-actions]

[Backlog Discovery]
Reviewer: Product Manager

• Review execution result is unavailable.
• Please rerun the Reviewer workflow.