Build Buy Or Duct-Tape Your Pipeline

Some broken pipelines only need duct tape and a ticket with a date on it.

When a pipeline breaks mid-sprint, I ask two questions before anyone picks a fix.

How urgent is it? Is someone blocked today, or is this a slow leak?

How reversible is the fix? Can you undo it in a week, or will it shape everything downstream for a year?

Put those on a grid:

• Urgent, easy to undo: duct-tape it. A retry or a manual backfill. Then open a ticket with a date on it.
• Urgent, hard to undo: patch now, and put the real decision in next week’s planning.
• Can wait, easy to undo: buy. A managed connector usually costs less than an engineer’s month.
• Can wait, hard to undo: build, slowly, with a design doc. This is where architecture time pays off.

The box I see go wrong most is the second one. The 2am fix quietly becomes the design, because nobody comes back to it.

Which box is your current pipeline problem in? Tell me and I’ll reply with how I’d approach it.