For example.
After an initial Batch failure, Jobs can be manually requeued. When desired Job state is achieved, then setting the state to "UpdateReporting" will invoke reprocessing of the final portion of Batch processing, assessment and notification.
Technically, this will involve creating an additional Consumer Daemon (UpdateReporting)
Although, there may be a way to piggyback off of "BatchReport" daemon.
One concern. Will we need to disable hourly Job level cleanup to ensure Job data remains??
Maybe we can keep cleanup and exclude data in a Batch Failed state.
Batch Queue state Update Reporting is used to reassess a Failed batch state after manual intervention. Batch state diagram is shown in following page:
https://github.com/CDLUC3/mrt-zk/blob/main/design/states.md
For example. After an initial Batch failure, Jobs can be manually requeued. When desired Job state is achieved, then setting the state to "UpdateReporting" will invoke reprocessing of the final portion of Batch processing, assessment and notification.
Technically, this will involve creating an additional Consumer Daemon (UpdateReporting) Although, there may be a way to piggyback off of "BatchReport" daemon.
One concern. Will we need to disable hourly Job level cleanup to ensure Job data remains?? Maybe we can keep cleanup and exclude data in a Batch Failed state.