Troubleshoot Events That Do Not Start a Run

Use this guide when an external event should have started a workflow but no run appeared.

Before you begin

You need access to workspace events. You also need the expected integration connection slug, event name, and workflow trigger.

Check whether Shipfox received the event

Open Settings → Events. Filter by source, event name, and time.

If the event is absent, the problem is before trigger matching:

• For GitHub, check that the integration connection is enabled. Check that the App can read the repository. Repeat the GitHub setup checks.
• For Sentry, check the organization integration connection. Send a safe test issue. Repeat the Sentry setup checks.
• For a custom webhook, inspect the sender's HTTP response and delivery ID. Repeat Create a custom webhook checks.
• If the integration connection is disabled, enable it and send a new event.

Repair or replace an expired provider installation. Do the same for an unauthorized installation. Shipfox cannot receive a new event until then.

If delivery succeeded but no workflow matched

Open the event and compare it with the synced trigger:

1. source equals the integration connection slug on the event.
2. event equals the delivered event name.
3. The filter reads fields that exist in this payload and returns true.
4. The workflow finished syncing before the event arrived.

A false filter creates no run. A filter error does the same. Fix the trigger and wait for a successful sync. Then send a new event. Shipfox does not replay the old event after an edit.

Use Trigger sources for exact event contracts and Expressions for filter behavior.

If evaluation or run creation failed

A Failed result means Shipfox received the event. It could not finish the check or create the matched work.

Read the error on the event. Check the workflow's latest sync result and saved trigger. Fix the item named by the error. It may be an expression, integration connection, or workflow field. Wait for sync, then send one new test event.

Each matching trigger creates its own run. If only some runs appear, check each missing trigger on its own.

If the event stays evaluating

Compare its time with newer events from the same workspace. If newer events finish while this one stays Evaluating, record:

• Event or delivery ID.
• Integration connection slug and event name.
• Received time.
• Expected workflow.
• The last visible routing state.

Send one fresh test event. If it also stays evaluating, stop changing the workflow. The routing work did not finish. The trigger policy is not the likely cause.

Confirm the repair

Send a new safe event and confirm one final outcome:

• Triggered links to every expected run.
• No workflows triggered is correct for a deliberate non-match.
• Failed and Evaluating no longer appear for the repaired path.