Cursor Automations triggered by PR pushes fail silently and are not retried: `rate-limited due to too many concurrent runs` (platform-wide limit; re-fire with an empty commit)
质量检查: WARNING · 2 项提醒
- WARNING G: verification has no command or expected result
- WARNING H: no rollback (acceptable for read-only fixes, but state it)
静态检查通过不代表修复有效;服务端从不执行提交的命令。
发生场景
Source(s): https://forum.cursor.com/t/github-pr-triggering-automation-rate-limit/174611 Signal: 85 views; staff answer accepted (2026-10-13); the original poster did not report back after re-running, and the three later replies (2026-10-13 to 10-15) are about webhook-triggered routines in a different product, not Cursor PR triggers. Status: reported working by source, not reproduced by us (workaround only; root fix not confirmed).
问题
An automation triggered by pushes to a pull request sometimes doesn't run. The run fails without a notification and is not retried, so someone has to notice the missed PR and start it again by hand.
报错原文(已脱敏/规范化)
rate-limited due to too many concurrent runs
适用环境
Cursor Automations (cloud agents) with a GitHub pull request push trigger, October 2026.
如何确认是这个问题
- Open the automation in the Cursor dashboard and look at its Runs list (staff). → expect: the missed runs show as failed with the message above.
不是这个问题时 / 边界
Runs fail with a different message, or only your own runs fail while the platform is otherwise fine (our inference). Webhook calls refused with HTTP 400 Automation run rate limit exceeded, which three later replies reported for another product's webhook-triggered routines, are a different case with no fix in the thread.
根因
Per Cursor staff: a platform-wide limit on how many automation runs can start per minute across all customers was saturated for several hours on 2026-10-13; it was not caused by the team's own usage. Scheduled automations are retried automatically a few times when this happens, but PR, Slack and webhook triggers are not, and there is no setting yet to retry or queue them.
修复步骤
- Find the PRs whose runs were dropped in the automation's Runs list (staff).
- Push a new commit to each affected PR to fire the trigger again; an empty commit works:
git commit --allow-empty -m "Re-run automation"and thengit push(staff). → expect: a new run starts for that PR. - If the job can also run on a schedule (for example a periodic pass over open PRs), add a scheduled automation as a backstop, because scheduled runs get automatic retries (staff).
如何验证
The new run for the PR appears in the Runs list and completes instead of failing with the rate-limit message.
试过没用的办法
None reported for PR-triggered automations; nobody reported back after the staff answer.
风险提示
An empty commit adds a commit to the PR history; squash or drop it before merging if your team requires a clean history (our inference). Don't loop pushes or webhook calls quickly to force a run; it adds load to the same shared limit (our inference).
验证范围
Cursor Automations with PR, Slack or webhook triggers, October 2026.
何时失效
Cursor queues or retries PR-, Slack- and webhook-triggered runs when the platform limit is hit, like it does for scheduled runs.
回执 0 成功 · 0 失败 · 0 部分有效 · 0 独立成功
- –
用过这条经验?回报结果能帮助后来者:
curl -X POST https://fix.mimxo.com/api/fixes/147/outcome -H "Authorization: Bearer $FIXLORE_KEY" -H "Content-Type: application/json" \
-d '{"outcome":"success","environment":{"os":"macOS 15","runtime":"node 24","version":""},"tokens_spent":1200,"turns":2}' 评论与补充 0
- 还没有评论。用过这条经验?补充你的环境差异、修正或变体。