Increased task-related errors for Copilot Mission Control
On November 4, 2026, between approximately 03:40 UTC and 10:01 UTC, some Copilot agent users experienced errors listing and creating agent tasks. On average, 1.22% of requests were failed, with a peak of 2.68% over a ten-minute interval.<br /><br />The issue occurred when increased task-list traffic and expensive queries exhausted shared database capacity. We mitigated the broader impact by disabling a client experiment and temporarily restricting task-list requests from the affected client. Task-list requests remained restricted after the wider service recovered; we removed the blanket restriction on November 6 and continued a gradual rollout.<br /><br />We’ve since improved task-list queries, moved rate limiting to separate infrastructure and added capacity monitoring. We’re continuing to reduce unnecessary client requests and validate capacity before expanding the rollout to reduce the chance of recurrence.
The degradation has been mitigated. We are monitoring to ensure stability.
We're continuing to see recovery as rollout continues
We've identified the issue with a specific client version and are deploying a mitigation. We are already seeing partial recovery.
Copilot Mission Control is experiencing increased error rates (approx 2.5%) in listing and creating tasks due to upstream database throttling. We are investigating two potential mitigations.
We are investigating reports of impacted performance for some GitHub services.
關於這些資料
GitHub 的事件資料來自哪裡?
資料來自GitHub 官方狀態頁,每 5 分鐘同步,保留最近 90 天。詳情頁提供官方記錄連結。
如何解讀元件狀態列?
每格代表一天。綠色表示無異常記錄,黃色表示部分降級,紅色表示較大故障,灰色表示無資料。可用率統計涵蓋最近 90 天。
為什麼有些服務查不到歷史?
只收錄提供公開官方狀態頁的服務。
持續時間如何計算?
按官方標記的開始至復原時間計算。進行中的事件計算至目前時間。跨天事件會顯示在每個受影響的日期。
完整的服務清單與即時狀態見狀態總覽