Ignore superseded dashboard publish cancellations

[trask]

Treat cancelled per-repository dashboard runs as non-failures for hourly backfill notification purposes, so superseded publish jobs do not open noisy workflow failure issues while real failures still report.

[Copilot]

Pull request overview

This PR makes a targeted one-line change to the hourly Pull Request Dashboard workflow's failure-notification gate. Previously, the notify-hourly-failure job treated only a success result from the per-repository run-repo-dashboard matrix job as a non-failure. Because the called publish-dashboard job uses concurrency with cancel-in-progress: true, a run that is superseded by a newer publish surfaces as cancelled, which was incorrectly flagged as a failure and opened noisy workflow-failure tracking issues. The change now also treats cancelled as a non-failure. Genuine failures still report, since matrix result aggregation reports failure in preference to cancelled.

Changes:

• Treat a cancelled result from run-repo-dashboard as success for hourly backfill failure notifications, alongside success.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.