We're using VMWare Esxi virtual machines with ElasticSearch 0.19.8 cluster
and have some troubleshooting to do.
Index size is about 50Mb, Virtual machines RAM 2Gb. Testing via
ApacheBenchmark tool gives 1-3ms response time with concurrency 1, 10-30ms
with concurrency 5, 100-300ms with concurrency 10. Concurrency more than 10
fully crashes the system.
How do we properly configure ElasticSearch to increase concurrencies?