How to clean up elasticsearch indices when elasticsearch fills up the disk of a data store in OSP's centralized logging solution?

Issue

As part of Red Hat OpenStack's tech preview, one can configure a centralized logging architecture.

After some time, elasticsearch is filling the whole disk of a data store in Red Hat OpenStack Platform's centralized logging solution.

[root@datastore ~]# sudo find / -size +100000 -exec ls -la '{}' 2> /dev/null \; | sort -k 5 -n| tail -n20
-rw-r--r--. 1 elasticsearch elasticsearch 246367839 Jun 21 17:41 /var/lib/elasticsearch/datastore/nodes/0/indices/logstash-2016.06.21/1/index/_aa2.fdt
-rw-r--r--. 1 elasticsearch elasticsearch 246937268 Jul  4 17:05 /var/lib/elasticsearch/datastore/nodes/0/indices/logstash-2016.07.04/4/index/_bbk.fdt
(...)

How can one clean up the disk?