As stated here, GitHub is experiencing issues with honouring scheduled jobs in workflows.
This is creating downstream issues for Pod authors: the CDN rebuild is now firing roughly every 100 min instead of every 5. For pods with internal dependencies, that turns a release into a multi-hour wait, since each dependent pod can't resolve until its dependency propagates to the CDN.
The stated workaround as of last week (from the link above and at the time of writing - 12th June '26):
The workaround that actually holds up is to ditch the cron trigger and use an external service to fire workflow_dispatch instead. cron-job.org is free. Point it at: https://api.github.com/repos/{owner}/{repo}/actions/workflows/{workflow_id}/dispatches
Build.yml already declares repository_dispatch: [cronjob], so an external scheduler could fire it today with no new trigger. The Deploy to Pages step is currently gated if: github.event_name == 'schedule', so that condition would need to also allow repository_dispatch
Is adapting to this workaround something that would be considered? If so, I'm happy to open a PR
As stated here, GitHub is experiencing issues with honouring scheduled jobs in workflows.
This is creating downstream issues for Pod authors: the CDN rebuild is now firing roughly every 100 min instead of every 5. For pods with internal dependencies, that turns a release into a multi-hour wait, since each dependent pod can't resolve until its dependency propagates to the CDN.
The stated workaround as of last week (from the link above and at the time of writing - 12th June '26):
The workaround that actually holds up is to ditch the cron trigger and use an external service to fire workflow_dispatch instead. cron-job.org is free. Point it at: https://api.github.com/repos/{owner}/{repo}/actions/workflows/{workflow_id}/dispatches
Build.yml already declares repository_dispatch: [cronjob], so an external scheduler could fire it today with no new trigger. The Deploy to Pages step is currently gated
if: github.event_name == 'schedule', so that condition would need to also allow repository_dispatchIs adapting to this workaround something that would be considered? If so, I'm happy to open a PR