If the need arises, you can restore the state from a specific snapshot using the following command:

curl -XPOST 'http://localhost:9200/_snapshot/backups/backup_201905271530/_restore'

This will restore the backup_201905271530 snapshot from the backups repository. 

Once we have set up a periodic job that takes and stores a snapshot, we are safe in the event of any failure. We now have a cluster that is recoverable from any disaster-like situation. Remember, the output of snapshots should be stored in resilient storage. At least, it should not be saved on the same Elasticsearch cluster; it should be saved on different storage, preferably a robust filesystem that is highly available, such as S3, HDFS, and so on.

So far in this chapter, we have got up and running with a cluster that is reliable and is backed up regularly. In the upcoming sections, we will see how to address some common scenarios in data modeling. We will see some common strategies for setting up aliases for indexes, index templates, modeling time-series data, and so on.