Actions 和 Checks API 故障
2026年11月9日协调世界时18:07至18:59期间,GitHub Actions出现性能下降。在此期间,部分客户遇到工作流运行无法启动、更新或正确报告状态的问题。平均28%的工作流运行受到影响,在事件最严重时段峰值达到65%。Actions服务恢复后,部分作业遗留的状态限制了可用并发数,直至运行被释放。<br /><br />该事件发生在将Actions核心数据库从GitHub数据中心迁移至新云区域的计划内迁移期间。迁移完成后,我们将Actions服务流量切换至新区域。请求量超出了该区域配置的容量和速率限制,因此部分请求被拒绝,工作流更新失败。包括Webhooks、Pages和Copilot代码审查在内的依赖服务受到了二次影响。<br /><br />我们通过将部分流量切回原区域来缓解影响,使服务流量保持在配置限制内。随后我们提高了新区域的容量和速率限制,再逐步将流量切回。虽然此次迁移是该集群的一次性事件且现已完成,但我们正在改进剩余迁移期间在切换流量前验证容量和限制的方式,包括以更小的步骤切换流量。为解决事件后遗留的卡住运行,我们正在改进未来检测和缓解此类运行的能力,并准备一个API使客户能够自行处理。
原文On November 9, 2026, between 18:07 and 18:59 UTC, GitHub Actions experienced degraded performance. During this time, some customers saw workflow runs fail to start, update or report status correctly. An average of 28% of workflow runs were affected, peaking at 65% during the worst period of the incident. Once Actions services recovered, some jobs were left in a state that limited available concurrency until the runs were released. <br /><br />The incident occurred during a planned migration of a core Actions database from GitHub’s datacenter to a new cloud region. After the migration completed, we shifted Actions service traffic to the new region. The volume of requests exceeded the capacity and rate limits configured there, so a portion of requests were rejected and workflow updates failed. Dependent services, including Webhooks, Pages and Copilot code review, saw secondary impact. <br /><br />We mitigated impact by shifting some of the traffic back to the original region, allowing service traffic to remain within the configured limits. We then increased capacity and rate limits in the new region before gradually shifting traffic back. While this was a one-time migration event for this cluster that is now complete, we are improving how we validate capacity and limits before shifting traffic during remaining migrations, including shifting traffic in smaller steps. To address the stuck runs left after the incident, we are improving our ability to detect and mitigate these runs for future cases as well as readying an API to enable customers to address this themselves.
影响 Actions、Pages 和 Webhooks 的降级问题已得到缓解。我们正在监控以确保稳定性。
原文The degradation affecting Actions, Pages and Webhooks has been mitigated. We are monitoring to ensure stability.
Pages 正在经历性能降级。我们正在继续调查。
原文Pages is experiencing degraded performance. We are continuing to investigate.
我们正在调查影响GitHub Actions和Checks API的错误率升高问题。一旦有更多信息,我们将分享进一步更新。
原文We're investigating elevated errors affecting GitHub Actions and the Checks API. We'll share another update as we learn more.
Webhooks 目前性能下降。我们正在继续调查。
原文Webhooks is experiencing degraded performance. We are continuing to investigate.
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.
我们正在调查有关Actions可用性下降的报告。
原文We are investigating reports of degraded availability for Actions
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