Add batch_size for slurm jobs

[normanrz]

Detailed Description

Some jobs that we run on a Slurm cluster (e.g. downsampling) take longer to queue than to actually run. Splitting it up in many small Slurm jobs is therefore inefficient. Just making bigger jobs (e.g. by increasing the chunks_per_shard) is also not a great option, because it increases the required RAM as well as storage units on disk. Therefore, it would be good to just schedule multiple jobs in one Slurm job. This would be configurable by the batch_size (or similarly named) job resources.