Going further

Record what to do when something fails

The moment you need this knowledge is 2 a.m. during an incident — exactly when the person who has it is asleep. Write down what breaks and what to do about it while it’s calm, and it’s there when it counts.

1. Select a node, open the Logic tab, and click + Add failure notes.

2. Write what fails here and what to do about it — "if this runs twice, the totals double; re-run job X to correct".

   A failure note attached to the object is found by whoever is staring at that object, not buried in a wiki nobody remembers.

3. Open the flow’s inspector to see the Failure runbook — every node’s failure note, gathered in flow order.

   Read top to bottom, it becomes an incident checklist for the whole flow: the known failure modes, in the order they happen.