背景:需求是这样的,原始文件是txt文件(每天300个文件),最终想要的结果是每天将txt中的数据加载到es中,开始的想法是通过logstash加载数据到es中,但是对logstash不太熟悉,不知道怎么讲程序弄成读取一个txt文件到es中以后,就将这个txt原始文件备份并且删除掉,然后就想到了通过一个中间件来做,Python读取成功一个txt文件,并且加载到kafka中以后,就将这个txt文件备份然后删除掉原始文件。
第一步:向kafka中添加数据,我用Python连接kafka集群,向其中加载数据
# -*- coding: utf-8 -*-import json
import json
import msgpack
from loguru import logger
from kafka import KafkaProducer
from kafka.errors import KafkaErrordef kfk_produce_1():"""发送 json 格式数据:return:"""producer = KafkaProducer(bootstrap_servers='192.168.85.109:9092',value_serializer=lambda v: json.dumps(v).encode('utf-8'))#logstash-topic-one#producer.send('python_test_topic', {'key': 'value'})producer.send('logstash-topic-one', {'name': 'value'})kfk_produce_1()
执行完的结果,来界面工具上看,显示这样,说明数据已经加载进来了
第二步:配置logstash,将kafka中的数据加载到es集群中
编写的logstash.conf配置如下;
input{kafka{bootstrap_servers => "192.168.85.109:9092"client_id => "consumer_id"group_id => "consumer_group"auto_offset_reset => "latest"consumer_threads => 1decorate_events => truetopics => ["logstash-topic-one","logstash-topic-two"]}
}
output {if [@metadata][kafka][topic] == "logstash-topic-one" {elasticsearch {hosts => "http://192.168.85.109:9200"index => "kafka-one-data"timeout => 300}}if [@metadata][kafka][topic] == "logstash-topic-two" {elasticsearch {hosts => "http://192.168.85.109:9200"index => "kafka-two-data"timeout => 300}}stdout {}
}
第三步:执行logstash,通过kibana查看数据是否在es集群中,展示如下,则说明配置是正确的
问题1:现在发现,name字段是在message下面,如果是多个字段的话,不方便查询,想着怎么讲字段从message中弄出来,修改的配置如下,增加一段这样的代码就OK了
type => "json"codec => json {charset => "UTF-8"}
完整的配置文件logstash.conf代码如下;
input{kafka{bootstrap_servers => "192.168.85.109:9092"client_id => "consumer_id"group_id => "consumer_group"auto_offset_reset => "latest"consumer_threads => 1decorate_events => truetopics => ["logstash-topic-one","logstash-topic-two"]type => "json"codec => json {charset => "UTF-8"}}
}
output {if [@metadata][kafka][topic] == "logstash-topic-one" {elasticsearch {hosts => "http://192.168.85.109:9200"index => "kafka-one-data"timeout => 300}}if [@metadata][kafka][topic] == "logstash-topic-two" {elasticsearch {hosts => "http://192.168.85.109:9200"index => "kafka-two-data"timeout => 300}}stdout {}
}
然后我又造了一个多字段的场景如下;
我先去logstash中查看日志如下,字段已经分离出来了
{"name" => "value","@version" => "1","type" => "json","@timestamp" => 2023-05-17T06:13:48.825Z
}
{"@version" => "1","type" => "json","@timestamp" => 2023-05-17T06:20:57.729Z,"name" => "令狐冲","age" => "30","height" => "180cm"
}
去kibana中去查询,显示如下,测试成功喽,😄
问题2:在查询结果中发现,有些字段是没有用的,看看怎么去掉?
在配置文件中增加一个过滤器就可以解决了
filter { mutate {remove_field => ["@version","@timestamp","type"] # 删除字段}}