New: Cookbooks and AI ExplanationsStep-by-Step recipes to solve problems connected to Roadmaps and Cheat Sheets. Need more details? Use AI buttons for structured and simple explanations with concrete examples throughout the whole platform.Take a look
The same incident keeps happening, and no fix ever sticks.
What you'll have at the end
A short RFC the team actually signs off on, with one committed follow-up action
You need
Evidence that one exact failure has already recurred more than once, with each past fix landing on a single pipeline instead of the shared cause behind it, and a rough idea for the one shared change that would actually stop it.
Not covered
Building the shared fix itself once the team agrees to it, or writing the after-the-fact record once it ships; this recipe stops at the point the team has signed off and named who builds the first piece.
Leans on
The same data postmortem keeps repeating, and the fix never sticks
start there first if you don't yet have proof this is the same failure recurring with a diagnosed shared cause; this recipe assumes that work is already done and the fix needs a shared redesign, not one more patched job.
Turn a hard-to-reverse platform choice into a written ADR
use that once this RFC is accepted and built, to leave the short permanent record of what was decided; this recipe stops at getting the redesign approved, not documenting it afterward.
Checked 26 Aug 2026
Part of the Data Engineering Core cookbook