Show HN: IncidentPost – 事故复盘，由 AI 自动生成，无需复杂工具
1 分•作者: ededft•6 个月前
Hi HN，
我开发了 IncidentPost (<a href="https://incidentpost.vect.pro" rel="nofollow">https://incidentpost.vect.pro</a>)，旨在帮助工程师和 SRE 们发布清晰、公开的事故复盘报告，避免繁琐的内部工具或手动编写的压力。许多团队在停机后，尤其是在时间紧迫的情况下，难以格式化时间线、总结根本原因，并生成可分享的报告。
它的功能：
* 粘贴原始时间线（Slack 日志、时间戳、警报）
* 生成结构化报告，包括执行摘要、根本原因分析、时间线、缓解措施和行动项
* 生成可立即发布的公开 URL、Markdown/HN 帖子草稿和 Twitter 线程
* 无需帐户或复杂的入职流程即可阅读公开报告——最大限度地降低透明度门槛
重要性：
* 公开事故复盘报告可以提高信任度，改进工程流程
* 将数小时的写作时间缩短为几分钟
* 专注于事故解决，而非文档编写
目前还处于早期阶段，以反馈为导向，并有意保持简洁。我非常欢迎处理可靠性和事故复盘的工程师、SRE 或创始人提出意见：
缺少什么？什么感觉是必要的？你会使用类似的东西吗？
查看原文
Hi HN,<p>I built IncidentPost (<a href="https://incidentpost.vect.pro" rel="nofollow">https://incidentpost.vect.pro</a>
) to help engineers and SREs ship clear, public incident postmortems without heavy internal tooling or manual writing stress. Many teams struggle to format timelines, summarize root causes, and produce shareable reports after outages — especially under time pressure.<p>What it does:<p>Paste a raw timeline (Slack logs, timestamps, alerts)<p>Generates a structured report with executive summary, root cause analysis, timeline, mitigations, and action items<p>Produces a public URL, Markdown/HN post draft, and Twitter thread you can publish instantly<p>No accounts or complex onboarding to read public reports — minimal barriers for transparency<p>Why this matters:<p>Public postmortems improve trust and engineering processes<p>Automates hours of writing into minutes<p>Keeps focus on incident resolution, not documentation<p>It’s early, feedback-driven, and intentionally minimal. I’d really appreciate thoughts from engineers, SREs, or founders who deal with reliability and postmortems:
What’s missing? What feels necessary? Would you use something like this?