Add a `storage_intensive` resource boundary attribute for workers

This has come up in https://gitlab.com/gitlab-org/gitlab/issues/31178#note_279807061 and gitlab-org/gitlab!24479 (comment 283533989)

Should we add a new resource boundary for jobs that use a lot of local storage?

For sidekiq, this means the same as a memory bound worker: since the storage on nodes is limited, we should run to many of these jobs at the same time since they could be racing for the space. It also means the job cannot be latency sensitive (moving data around takes time).

If we decide to add this boundary, how many MB would a job be allowed to write before having to add this?