Table of Contents

Motivation

Problem 1:

Flink can be set up with retained checkpoints, leaving the last checkpoint when execution is cancelled or terminally fails. New jobs can resume from that checkpoint, left by a previous job.

...