Skip to content

Rerun Strategy

An approach that defines how failed or incomplete data jobs should be rerun and under what safety rules.

A rerun strategy is a foundational design decision for the operational resilience of data pipelines. It determines whether a failed job should be rerun fully, partially, or at a specific partition level. Poorly designed rerun logic can create duplicates or incomplete remediation. A strong rerun strategy should operate together with idempotency and lineage visibility.