Git 操作和拉取请求发生故障
2026年11月10日,大约从15:00 UTC到17:55 UTC,GitHub经历了性能下降,影响了仓库页面、拉取请求和GitHub Actions。客户可能遇到页面加载缓慢或失败,以及处理拉取请求更新和合并时出现延迟。在更广泛的服务降级结束后,一些拉取请求仍继续显示较早的源分支提交。GitHub Actions工作流约有1.8%的事件出现延迟,在16:30至16:50 UTC之间峰值约为7%。延迟通常为1.5至3.5分钟,最长可达10分钟。约0.09%的拉取请求事件未能启动其工作流,需要重新触发。<br /><br />此次降级源于支持Git操作的共享基础设施过载。近期的更改增加了后台处理以及计算拉取请求合并所需的工作量,给面向客户请求所使用的同一系统增加了压力。我们通过禁用近期的一项合并处理优化、暂停后台分析以及临时限制较低优先级流量来缓解此事件。临时限制还限制了未登录用户对提交页面的访问。服务健康恢复后,后台处理继续追赶进度。<br /><br />我们已减少重复的仓库分析工作,并正在添加工作负载级别的并发控制,以更好地保护面向客户的请求。我们还在改进安全降低后台负载的流程,并使拉取请求同步对临时故障更具弹性。
原文On November 10, 2026, from approximately 15:00 UTC to 17:55 UTC, GitHub experienced degraded performance affecting repository pages, pull requests, and GitHub Actions. Customers may have seen slow or failed page loads and delays processing pull request updates and merges. Some pull requests continued to display an earlier source-branch commit after the broader service degradation ended. GitHub Actions workflows were delayed for approximately 1.8% of events, peaking at about 7% between 16:30 and 16:50 UTC. Delays were typically 1.5 to 3.5 minutes, up to 10 minutes. Approximately 0.09% of pull request events did not start their workflows and needed to be re-triggered. <br /><br />The degradation resulted from overload in the shared infrastructure that supports Git operations. Recent changes increased background processing and the work required to compute pull request merges, adding pressure to the same systems used by customer-facing requests. We mitigated the incident by disabling a recent merge-processing optimization, pausing background analysis, and temporarily limiting lower-priority traffic. The temporary limits also restricted access to commit pages for users who were not signed in. Background processing continued to catch up after service health recovered. <br /><br />We have reduced repeated repository-analysis work and are adding workload-level concurrency controls to better protect customer-facing requests. We are also improving our procedures for safely reducing background load and making pull request synchronization more resilient to temporary failures.
我们继续调查Git操作和拉取请求响应时间升高的问题。Actions工作流正在恢复。拉取请求处理已显示出一些改善,但用户仍遇到延迟和失败。我们正在监控影响并努力恢复正常服务。
原文We continue to investigate elevated response times for Git operations and pull requests. Actions workflows are recovering. Pull request processing has shown some improvement, but users continue to experience delays and failures. We’re monitoring the impact and working to restore normal service.
Actions 运行正常。
原文Actions is operating normally.
Pull Requests 正经历性能下降。我们正在继续调查。
原文Pull Requests is experiencing degraded performance. We are continuing to investigate.
我们正在继续调查Git操作和拉取请求响应时间升高的问题。部分拉取请求处理和Actions工作流正在改善,但延迟和失败仍然存在。我们正在监控影响并努力恢复正常服务。
原文We continue to investigate elevated response times for Git operations and pull requests. Some pull-request processing and Actions workflows are improving, but delays and failures persist. We're monitoring the impact and working to restore normal service.
Actions 正经历可用性降级。我们正在继续调查。
原文Actions is experiencing degraded availability. We are continuing to investigate.
我们正在调查 Git 操作和拉取请求响应时间升高的问题。由拉取请求触发的 GitHub Actions 工作流出现延迟。我们正在努力减轻影响,并将在了解更多信息后分享更新。
原文We're investigating elevated response times for Git operations and pull requests. GitHub Actions workflows that are triggered from Pull Requests are delayed. We're working to mitigate the impact and will share updates as we learn more.
Actions 和 Pull Requests 目前可用性下降。我们正在继续调查。
原文Actions and Pull Requests are experiencing degraded availability. We are continuing to investigate.
我们正在调查有关Actions和Pull Requests性能下降的报告。
原文We are investigating reports of degraded performance for Actions and Pull Requests
关于这些数据
GitHub的故障事件数据从哪来?
全部来自GitHub 官方状态页的公开接口,由本站每 5 分钟同步一次,保留最近 90 天。事件标题、时间、影响级别与更新记录均为官方原文,本站不做改写;点详情页底部的链接可回到厂商原始记录核对。
顶部的组件状态条怎么看?
每一格是一天,绿色表示GitHub 官方状态页当天未记录异常,黄色表示部分降级,红色表示当天有较大故障,灰色表示该组件当时还没有数据。右侧的可用率是官方口径的 90 天统计,与状态总览页显示的是同一份数据。
为什么有的服务查不到历史?
本页只收录对外提供官方状态页的服务。没有公开状态页的厂商(多数国产大模型属于此类)无法取得可信数据,本站不做自建拨测去猜,因此也不会出现在状态总览里。
持续时长怎么算?
按官方标注的开始时间到恢复时间计算;尚未恢复的事件按「至今」计算并标为进行中。跨天的事件在状态条上会覆盖它经过的每一天。
完整的服务清单与实时状态见状态总览