Pause the execution of a batched background migration when too many jobs fail

Context:

Currently, when we have a bug/issue on a batched background migration, the system executes the whole migration and ignores that a significant number of jobs are failing. We don't have any autonomous mechanism to stop the migration in case of a problem. Usually, we do this manually (with human interaction).

Solution:

We should stop the batched background migration when a certain number of jobs fail.