Issue
When using a Webhook trigger (Respond = Using 'Respond to Webhook' Node) followed by a GitHub node set to Dispatch and Wait for Completion, the webhook connection is dropped as soon as the GitHub node enters wait mode. The workflow later resumes and reaches Respond to Webhook, but the original caller has already disconnected and never gets the response.
Repro steps
- Workflow:
- Webhook (keep open, respond via Respond to Webhook node)
- GitHub → Operation: Dispatch and Wait for Completion
- Respond to Webhook
- Call the production webhook URL.
- When GitHub enters wait, the HTTP connection is closed.
- On resume, Respond to Webhook executes, but no response is delivered (caller timed out).
Expected
The webhook connection should stay open until Respond to Webhook returns after the wait/resume cycle.
Notes
- Works fine if Webhook → Wait (external call) → Respond to Webhook.
- Breaks only when a Dispatch and Wait node (or any intermediate node before wait) is used.
- Likely the webhook context is lost when execution is persisted for wait.
Question
Is this expected behavior, or a bug with how n8n handles waits when combined with webhook responses?