ES:6.3.2
Logstash: 6.3.2-1
input{
    kafka{
       bootstrap_servers => 'kafka-end-point:9092'
       topics => ["topic1", "topic2", "topic3", "topic4","topic5","topic6","topic7","topic8"]
       codec => "json"
       group_id => "logstash-ec2-2"
       consumer_threads => "60"
       decorate_events => true
    }
    }
    filter {
            json {
             source => "message"
             #target => "appmessage"
            }
            json {
             source => "msg"
             #target => "[message][atts]"
      }
    }
    output{
                    elasticsearch {
                           hosts => ["http://es-node:9200"]
                           index => "%{[@metadata][kafka][topic]}-%{+YYYY.MM.dd}"
                           codec => "json"
                           sniffing => false
                    }
    }
It runs for some time without issues and then stops reading logs from Kafka
We get this error
    [2018-10-25T12:17:08,405][ERROR][logstash.outputs.elasticsearch] Attempted to send a bulk request to elasticsearch' but Elasticsearch appears to be unreachable or down! {:error_message=>"Elasticsearch Unreachable: [http://es-node:9200/][Manticore::ClientProtocolException] es-nodes.production.pm.com:9200 failed to respond", :class=>"LogStash::Outputs::ElasticSearch::HttpClient::Pool::HostUnreachableError", :will_retry_in_seconds=>2}
[2018-10-25T12:17:08,428][WARN ][logstash.outputs.elasticsearch] UNEXPECTED POOL ERROR {:e=>#<LogStash::Outputs::ElasticSearch::HttpClient::Pool::NoConnectionAvailableError: No Available connections>}
I have verified that elasticsearch is accessible from logstash machines
We have 13 nodes of ES fronted by an ELB(AWS). All of them are healthy