My elasticsearch cluster contains two nodes: node_ip1,node_ip2.
When I used logstash-output-elasticsearch to connect elasticsearch cluster, there were lots of tcp time_wait between logstash and elasticsearch nodes. Here is my logstash configuration:
output {
elasticsearch {
hosts => ["http://node_ip1:9200"]
manage_template => true
index => "logstash-%{+YYY.MM.dd}"
flush_size => 1500
idle_flush_time => 10
workers => 4
codec => "json"
}
}
As above, logstash agent had connected to node_ip1:9200, but there were lots of tcp time_wait between logstash agent and node_ip2:9200 (Another elasticsearch node).
How do I resolve this problem?