Scheduled workflow has no run records for multiple expected cron slots (Asia/Shanghai) #212997
🏷️ Discussion TypeBug 💬 Feature/Topic AreaSchedule & Cron Jobs Discussion DetailsSummaryI am seeing intermittent missing scheduled-run records in an active public repository. Some scheduled runs occur, while several expected slots have no corresponding run visible in the Actions UI or REST API at the observation time. I am asking for help distinguishing normal delay/dropped events from a scheduler eligibility or registration problem; I am not claiming a confirmed backend root cause.
Relevant configurationon:
schedule:
- cron: "0 0,12,18 * * *"
timezone: Asia/Shanghai
- cron: "7,27,47 * * * *"
timezone: Asia/Shanghai
workflow_dispatch:
inputs:
archive_only:
type: boolean
default: false
concurrency:
group: tariff-scheduled-refresh
cancel-in-progress: falseThe first schedule is intended for 00:00, 12:00 and 18:00 Beijing time. The second is a catch-up check: its plan job exits without collection if the current slot has already completed. That job-level skip is expected, but should still leave a workflow run record. The issue here is the absence of a run record, rather than a collector step failing or intentionally skipping. There is also a push trigger limited to .github/collection-request.json for an explicit catch-up request. Concrete observation on 2026-11-13By 04:49:06 UTC (12:49:06 Asia/Shanghai), I could not find a schedule-triggered run for any of these expected UTC slots:
The last slot was only about two minutes old when checked, so its absence alone is not evidence of a dropped event. The main noon slot was about 49 minutes old. These are observation-window results, not proof that an event could never arrive later. Other schedule-triggered runs do exist with the same schedule configuration:
An explicit push-triggered catch-up run was created at 04:49:40 UTC, started normally, and completed collection/archive: Checks already performed
I have read the schedule documentation: delays and dropped events can occur under load, especially at the start of the hour. This is why I added checks at minutes 7, 27 and 47, and why I do not treat absence as proof of a particular cause. I also searched Community for similar schedule-not-triggered discussions before posting. GitHub Support referred this account to self-service/community resources without providing scheduler diagnostics. Questions
Only public repository links and run identifiers are included; no credentials or webhook addresses are needed for this investigation. |
Replies: 2 comments
|
💬 Your Product Feedback Has Been Submitted 🎉 Thank you for taking the time to share your insights with us! Your feedback is invaluable as we build a better GitHub experience for all our users. Here's what you can expect moving forward ⏩
Where to look to see what's shipping 👀
What you can do in the meantime 💻
As a member of the GitHub community, your participation is essential. While we can't promise that every suggestion will be implemented, we want to emphasize that your feedback is instrumental in guiding our decisions and priorities. Thank you once again for your contribution to making GitHub even better! We're grateful for your ongoing support and collaboration in shaping the future of our platform. ⭐ |
|
Hi @JayChou0609, Your checks already cover most of the usual configuration issues. GitHub's documentation confirms that scheduled workflows can be delayed or dropped during periods of high load, especially at the start of an hour. Since your workflow is active and other scheduled runs are working, I wouldn't assume the YAML is incorrect. For recovery, one option is to use an external scheduler that checks whether the daily collection completed. If it hasn't, the scheduler could trigger your existing workflow_dispatch through the GitHub API. You already have a catch-up mechanism, so this could work alongside it without depending entirely on GitHub's cron schedule. Documentation: This wouldn't identify why the original event was missed, but it could make the daily collection more reliable. |
Hi @JayChou0609,
Your checks already cover most of the usual configuration issues.
GitHub's documentation confirms that scheduled workflows can be delayed or dropped during periods of high load, especially at the start of an hour.
Since your workflow is active and other scheduled runs are working, I wouldn't assume the YAML is incorrect.
For recovery, one option is to use an external scheduler that checks whether the daily collection completed. If it hasn't, the scheduler could trigger your existing workflow_dispatch through the GitHub API.
You already have a catch-up mechanism, so this could work alongside it without depending entirely on GitHub's cron schedule.
Documentation:
https://do…