n8n Incident Alerts: Airtable, OpenAI & Optional Slack

n8n Incident Alerts: Airtable, OpenAI & Optional Slack—turn workflow failures into actionable incidents

When your n8n automation workflow errors, this n8n Incident Alerts workflow captures the failure event, deduplicates noisy repeats, uses OpenAI to classify what happened, and stores incident state in Airtable—with optional escalation to Slack (and Gmail) plus an optional controlled retry.

What this workflow does

• Receives incident events via a webhook (or a manual test trigger) and normalizes key fields such as workflow ID, failed node, error type, and message.
• Deduplicates alert noise by querying Airtable for recent incidents matching the same workflow, error type, and failed node, then suppressing duplicates by updating the most recent record within the dedupe window.
• Classifies incidents with OpenAI by sending the normalized details to OpenAI to return concise JSON containing a summary, category, and a retry recommendation.
• Computes a deterministic severity using business impact/error signals and the OpenAI outputs, then creates an “Open” incident record in Airtable.
• Optionally triggers a controlled retry when the incident is low/medium, marked retryable, and OpenAI recommends retrying. It calls a protected HTTP retry endpoint and updates Airtable to “Retrying” when the retry starts successfully.
• Escalates when needed if retry isn’t safe or fails to start: it prepares a human escalation message and can post it to Slack and/or email it via Gmail. It returns a JSON result to the webhook caller when applicable.
• Daily unresolved summary runs on a schedule to load unresolved incidents from Airtable and generate a daily view of what still needs attention.

Use cases

• Detecting repeated n8n execution failures and preventing “alert spam” with Airtable-based deduplication.
• Automatically triaging incident severity and category using OpenAI for faster incident response.
• Retrying transient errors safely via a protected HTTP endpoint, then updating Airtable incident status.
• Escalating only what matters to Slack and/or Gmail for operational visibility.

Technical details (integrations & nodes)

• Airtable for incident state, dedupe updates, and tracking “Open”/“Retrying”/unresolved items
• OpenAI for JSON incident classification (summary, category, retry recommendation)
• Slack (optional) for escalation messages
• Gmail (optional) for email notifications
• n8n nodes including: if, set, code, no op, gmail, slack, plus webhook/manual trigger and scheduled run