This topic was automatically generated from Slack. You can find the original thread here.
Outlook triggers deployed via Connect die silently: Graph subscription renewal 404s and never re-creates the subscription
We use Pipedream Connect to deploy microsoft_outlook-new-email triggers (v0.1.19) on behalf of our end users.
What we’re seeing: A deployed trigger received events normally until Jul 15, 2026 ~19:02 UTC, then went quiet. The component’s built-in webhook renewal timer (every 2 days, since Graph subscriptions expire in ~3 days) keeps firing, but every run since then fails the same way:
• 2026-07-17T08:18 and 2026-07-19T08:18 (UTC): db.get hookId => "<hook-id>" followed by Error: The object was not found.
Our read: the stored Graph subscription no longer exists on Microsoft’s side, so the renewal call 404s. The source just throws and retries the same doomed renewal every 2 days. It never falls back to creating a fresh subscription, so the trigger is permanently dead while still reporting active: true via the Connect API, and nothing surfaces to us programmatically.
Questions:
- Is this a known issue with the
microsoft_outlooksources? - Shouldn’t the renewal handler treat a 404 as “subscription gone, re-create and store the new hookId” instead of failing forever?
- Is there a recommended way to detect dead triggers via the Connect API today? We have ~26 Outlook email triggers deployed for users in this project and need them to self-heal.
Happy to share more trigger IDs and full logs.