Data Pipeline Reliability

Use this skill when the main question is whether a pipeline can be rerun, retried, backfilled, or recovered without corrupting data.

Boundaries

• Use data-engineering when the main decision is platform shape, marts, semantic metrics, or serving layers.
• Use data-architecture-strategy when the main decision is which architecture family to adopt.
• Use data-quality when the main concern is whether outputs are trustworthy once they are produced.

Default procedure