The Kubernetes project has announced that the Job success policy has graduated to General Availability as part of the v1.33 release. In batch workloads, users may want to mark a Job as succeeded even if some followers fail. The Job success policy allows users to specify early exit criteria using the .spec.successPolicy field. This field is especially valuable for scientific simulation, AI/ML, and High-Performance Computing batch workloads. Users can specify a minimal required size of succeeded indexes for a Job to be marked as succeeded. Once a Job meets the success policy, it is marked as succeeded, and all Pods are terminated. The success policy can be customized to specify a specific index or count of indexes that must succeed for the Job to be marked as succeeded. The Job controller adds a SuccessCriteriaMet condition to the Job status once the Job reaches the success policy, and then initiates cleanup and termination of Pods. The Job obtains a Complete condition once the cleanup and termination are finished.