Shard Failing : queue error


Am getting this error and seems like can it be resolved by increasing queue size. Is my assumption correct ?

[2015-10-19 20:41:23,495][DEBUG][ ] [Ghost Rider 2099] [logstash-2015.10.13][4], node[gdLxmHjmSiiHqWBb6hpZhA], [P], s[STARTED]: Failed to execute [] lastShard [true]
org.elasticsearch.common.util.concurrent.EsRejectedExecutionException: rejected execution (queue capacity 1000) on$23@309f21ac

Where can I find the config file so as to change the parameters. Looked into elasticsearch.yml but couldnt find any such setings

You can temporarily relieve this by changing threadpools, but you are overloading the node so it'll probably happen again.

Whats the best way to overcome this problem ? Increasing nodes ? Is there a formula which tells me what is the best configuration to overcome this problem

There's no formula as it really depends on use case.

How are you monitoring ES?

I tried using marvel but it was quiet resource intensive

Also the indexer, is on a machine which is also running elasticsearch and kibana hence may be down on resources

What specs does the node have?

Its a Windows 2008 server with 4GB RAM. Though am trying to double the spec up but dont know if that would help to solve this problem

That's pretty small, adding more memory would help.