I am asking this more from a library standpoint - Is SolrJ, the java library for Solr the recommended way to work with Solr/SolrCloud in production systems? Or is it advised to work with simple HTTP requests? Also, how mature is the SolrJ project specially in relation to ingesting raw data to a new Solr/SolrCloud instance? Can I go ahead and use SolrJ to ingest data as well?
Yes, SolrJ is the preferred way of interacting with Solr in Java. It's also used internally by Solr for internode communication, and is the official library released together with Solr. You can't really get more suggested than that.
It should be able to do anything you like, and if there is no "native" support for the exact thing you want to do, you can make SolrJ do it through custom parameters.