If we move the elasticsearch logging cluster to span multiple machines, we'll probably want its configuration to be managed by puppet.
https://forge.puppet.com/elastic/elasticsearch/readme looks like a reasonable starting point (maintained by upstream, frequent releases).