Ceph » cephsqlite

Suggested by Josh in: https://github.com/ceph/ceph/pull/39191#discussion_r593526497

may want a way to limit parallelism in the future, like rbd does with rbd_concurrent_management_ops. not sure if this comes up in a sqlite workload, but if there are a large amount of i/o kicked off at once (like deleting a large db) it could take up a lot of cluster resources for an operation that may not be latency sensitive.