PatchRunner

PatchRunner is a web app for safely applying “hotfix” transformations to ETL jobs when upstream data breaks unexpectedly (schema drift, bad batches, null explosions). Instead of redeploying an entire pipeline or waiting for a full engineering cycle, teams can create a scoped patch: filter a bad partition, coerce a column type, quarantine suspect rows, or reroute to a fallback table. Patches are versioned, time-bounded, and auditable, with approvals and automatic rollback. It integrates with common orchestrators and warehouses so you can attach a patch to a specific job run or dataset. This is not a magic AI autopilot—most fixes are simple, but the value is speed, governance, and reduced downtime. Optional AI assistance can suggest patch templates based on error logs, but humans remain in control.