just like replication workers, allow failed replication workers to be configurable in situations like DR failures etc to catch up on replication sooner when DR is back online. Signed-off-by: Harshavardhana <harsha@minio.io>