It has been nearly two years since AWS Elastic Search was introduced. Microsoft Azure has however has still dragged its feet on any kind of Platform as a Service offering to handle Full Text Search work loads. That is to this day users will have to manage their own Elastic Search or Solr clusters. This of course not a painless process. Users would have to know and worry about the replication model and advanced concepts such as Data Nodes vs Edge Nodes vs Index Nodes vs Tribe Nodes. The following document by Azure describes how complex the topology of Elastic Search could potentially be.
There is of course the cloud offering by Elastic Search itself. however, when one is running a complex application cluster it would help to know that your full-text search nodes are managed by the same set of tools as the rest of your infrastructure. This is very helpful when one is planning regions pairings or which availability zones data is located in in. It is also the added the value that you manage your full-text search exactly with same management APIs as the rest of your environment. Any kind of enterprise application these days would have a technology such Elastic Search integrated as a vital component of that application block.