Keep the lead. Recover the alert.

The full kit's durable queue remembers each accepted inquiry and which delivery step finished. The free sample makes no external writes.

1. Inspect the queue

Use the included queue-admin tool to list jobs in your private installation. Record the leadId, stage, state and reason. Pending jobs retry automatically when due; failed jobs reached a retry limit or need a configuration fix; review jobs have an uncertain outcome. A successful n8n execution can still record a rejected delivery, so inspect the queue status.

2. Stop the old worker and check the destination

Before manually recovering a job, ensure the previous n8n execution has ended. Search Google Sheets by its exact leadId. Search the intended Slack channel for the Lead ID included in the alert. Do not rely on email alone: a different inquiry from the same person can be legitimate.

3. Acknowledge or retry only the held step

If the current step already happened, use queue-admin's delivered action with the lead ID and a private review note. Acknowledging Sheets advances the job to Slack; acknowledging Slack completes it. If the action definitely did not happen, fix the connection and use retry with a note. This resets the attempt allowance for that stage. Never clear the database or rerun only an external node to bypass the queue.

4. Understand automatic retries

There are at most five attempts per stage, with default waits of 30 seconds, 2 minutes, 10 minutes and 30 minutes. Rate limits respect Retry-After with at least a one-minute pause. Temporary Sheets server failures retry using the same leadId. Slack server failures and network timeouts are held because their outcome may be uncertain. The timer checks every 10 seconds while the workflow is published.

5. Keep duplicate protection intact

Identical normalized inquiries within 24 hours reuse the original lead ID. Unresolved inquiries remain protected beyond that window. A changed message or other field is a distinct inquiry. Keep one authoritative queue and one writer for the managed sheet. Google Sheets and Slack do not provide an end-to-end exactly-once guarantee.

Backups and retention

Back up the queue's persistent data and private key together with your n8n database and encryption configuration. After restoring an older backup, pause delivery and reconcile actions completed since the backup before resuming. Completed queue records are retained for 30 days; unresolved records remain until reviewed. Define your own retention policy across the queue, n8n, Sheets and Slack.

Before real use

The full kit requires self-hosted n8n and its included private queue companion. Keep the editor and queue private, add suitable form abuse controls and review held jobs daily. Test real account access and delivery in your own environment. No outgoing failure-notification service is included.