"Ignore" is the default error handler in Make. It means: if something fails, skip it and continue. In production, this silently drops data. A lead form submission fails to create an Airtable record? Ignored. The client never finds out until they notice a gap in their data weeks later.
Level 1: Break + Rollback
Set the error handler to "Break" for critical modules. This stops the scenario and marks the incomplete execution in Make's history. You can reprocess it manually once the issue is fixed.
Use Rollback where your scenario modifies multiple records — it reverts all changes in the current execution to keep data consistent.
Level 2: Error route in the scenario
Add an error-handling route (right-click a module → Add error handler). On failure, branch into a Telegram or Slack notification with the error details. The main flow breaks, but you get an immediate alert.
Level 3: Dedicated error-handler scenario
For critical flows, build a separate "error handler" scenario triggered by a webhook. Your main scenario catches errors and calls this webhook with context (scenario name, module, error message, input data). The handler logs to Airtable and sends a Telegram alert with a direct link to the failed execution.
This is the pattern I use for anything client-facing. If an order bot fails at 2 AM, I know about it at 2 AM — not the next morning when the client messages.