Notification node

Output & Observability

The Notification node sends an alert to a configured channel (email, Slack, Teams, generic webhook) from inside a workflow. Every row that enters the node triggers one notification, with the message text rendered from a per-row template.

When to use

• Downstream operators need to know a run happened, succeeded, failed, or emitted a specific row.
• You want to alert on a filtered subset — chain Filter into Notification so only rows matching a rule notify.
• You want notification bodies to include row data via templated placeholders.

Ports

Port	Direction	Kind	Notes
in	in	data	Rows to notify on.
out	out	data	Pass-through of the input rows so downstream nodes see the same rows the notification saw.
error	out	control	Fires on channel adapter errors (SMTP down, Slack rate limit, etc.).
notify	out	control	Meta-notification port for chained notification policies.

Configuration

• Channel. Email, Slack, Teams, or Generic webhook. The choice drives which additional fields appear.
• Recipient. Address column or literal (email address, Slack channel id, Teams webhook URL, webhook URL).
• Subject / message template. {{ row.column }} and {{ exec.varName }} placeholders are substituted per row.
• Batching.
  • per_row (default) — one notification per input row.
  • per_batch — one notification per input chunk (or per run, in the finalize pass).
• On failure.
  • route_to_error (default) — a channel error routes the row to error; the run continues.
  • retry_then_error — retry the channel a bounded number of times before erroring.
  • silent — swallow the error and drop the row.

Configure Action walkthrough

1. Pick the channel; the modal reveals the channel-specific fields (SMTP recipient, Slack channel, Teams webhook, etc.).
2. Pick the recipient column (or type a literal address).
3. Write the subject / message templates. Preview substitutes with the upstream sample.
4. Toggle batching mode. Choose failure policy.

Runtime behaviour

• Per-row notifications fire synchronously; per-batch fires once per chunk / once at finalize (for once_at_end variants).
• Channel adapters share the runtime's connector cache; SMTP connections are pooled across runs.

Failure modes

• Channel down. Routes to error under the default failure policy. Retry policy kicks in when configured.
• Recipient invalid. Adapter-specific error surface: SMTP returns 550, Slack returns channel_not_found, Teams returns the webhook error string.

Related nodes

Log

Emit an observability event without sending a channel message

Approval

Chain a notification off the approval events

Filter

Only notify on the interesting rows