Skip to main content

Recover a failed run automatically

What this solves

When a job fails in the middle of the night, two things usually go wrong: the mess it leaves (a half-written file, a lock, a partial post) sits there until morning, and nobody finds out until a business process is already late. The run just stops, and recovery waits for a person who isn't watching.

What you get. The workflow cleans up and retries the recoverable failures on its own, and raises a clear alert the moment something needs a human, so the run keeps moving and an operator only steps in for the failures that genuinely need judgment.