I am running elasticsearch 0.19.9 cluster with 36 instances.
Data pipeline has several dozen of TransportClient instances and I found
that only one client is getting actionGet timeout exception. I set the
timeout as 10s.
I checked 'ulimit -u' as 57408, so it's not the problem regarding the
number of threads. After I restarted the client, I am not seeing timeout
What should I have traced before restarting?