使用codec的multiline插件实现多行匹配，这是一个可以将多行进行合并的插件，而且可以使用what指定将匹配到的行与前面的行合并还是和后面的行合并。
1.java日志收集测试

input {stdin {codec => multiline {pattern => "^\["                        //以"["开头进行正则匹配negate => true                          //正则匹配成功what => "previous"                      //和前面的内容进行合并}}
}
output {stdout {codec => rubydebug }
}

2.查看elasticsearch日志,已"["开头

# cat /var/log/elasticsearch/cluster.log
[2018-05-29T08:00:03,068][INFO ][o.e.c.m.MetaDataCreateIndexService] [node-1] [systemlog-2018.05.29] creating index, cause [auto(bulk api)], templates [], shards [5]/[1], mappings []
[2018-05-29T08:00:03,192][INFO ][o.e.c.m.MetaDataMappingService] [node-1] [systemlog-2018.05.29/DCO-zNOHQL2sgE4lS_Se7g] create_mapping [system]
[2018-05-29T11:29:31,145][INFO ][o.e.c.m.MetaDataCreateIndexService] [node-1] [securelog-2018.05.29] creating index, cause [auto(bulk api)], templates [], shards [5]/[1], mappings []
[2018-05-29T11:29:31,225][INFO ][o.e.c.m.MetaDataMappingService] [node-1] [securelog-2018.05.29/ABd4qrCATYq3YLYUqXe3uA] create_mapping [secure]

3.配置logstash

#vim /etc/logstash/conf.d/java.conf
input {file {path => "/var/log/elasticsearch/cluster.log"type => "elk-java-log"start_position => "beginning"stat_interval => "2"codec => multiline {pattern => "^\["negate => truewhat => "previous"}}
}
output {if [type] == "elk-java-log" {elasticsearch {hosts => ["192.168.1.31:9200"]index => "elk-java-log-%{+YYYY.MM.dd}"}}
}

4.启动

logstash -f /etc/logstash/conf.d/java.conf -t
systemctl restart logstash 

5.head插件查看

6.kibana添加日志