The current cloud plugin creates several containers. One for the cluster
metadata, and 2 for each shard. This was done because of a limit that no
longer exists in S3 regarding the number of files per bucket. Because there
is no limit anymore, and there is still a limit of the number of buckets
created, I just pushed a change that stores everything under the same
container. A specific index can still be configured to be stored on a
As a side note, the limit on file size (S3 is 5G) is handled by
elasticsearch, since it will break files into chunks of 4G automatically.
Sadly, with this change, a full reindex will be required and creation of a
new container. I will release 0.7.2 soon with this change.