🏷️ Discussion Type

Bug

💬 Feature/Topic Area

ARC (Actions Runner Controller)

Discussion Details

TL;DR: schedule events were fully suppressed on my account (org Ning-Health) for ~3+ weeks, silently resumed 2026-10-02 hours after I filed ticket #4816841 (still never answered), and the suppression has now PARTIALLY RETURNED: my probe repo still receives its */5 cron, while my main repo's identical */5 cron has received zero schedule events for 14 days. Support has never replied with a human answer. Looking for (1) what account-level state causes this, (2) how to detect/reset it, (3) what diagnostics to collect.

Account/org: Ning-Health (GitHub Pro, upgraded during this saga). Issue first observed before 2026-09-23.

Timeline (all API-verified):

What makes the current state especially strange: two repos, same account, both active, both with */5 cron workflows — one is being scheduled right now, the other hasn't been scheduled in 14 days. Billing is clean (<$0.01 of included minutes), no platform incident covers the window, and workflow_dispatch/repository_dispatch continue to execute within seconds on the silent repo.

Questions:

1. Is there a per-repo or per-account scheduler state (shadow flag, queue detachment) that can suppress schedule delivery independently per repo, and can a user inspect or reset it?
2. What diagnostics can I collect from my side that would help GitHub staff identify the cause? (I can grant access to the probe repo, run controlled windows, capture API run-list diffs.)
3. Is partial recovery (one repo scheduled, another ignored) a documented failure mode?

The probe repo is left enabled with its */5 cron for direct verification. Happy to run any controlled test.