使用logstash收集javanginx系统等常见日志

Posted 2022-05-17 chuangcc

 

目录

1、使用codec的multiline插件收集java日志... 1

2、收集nginx日志... 2

3、收集系统syslog日志... 3

4、使用fliter的grok模块收集mysql日志... 4

 

 

1、使用codec的multiline插件收集java日志

对于采用ELK作为应用日志来说，多行消息的友好展示是必不可少的，否则ELK的价值就大大打折了。要正确的处理多行消息，需使用multiline插件

 

比如，对于java日志而言，可以使用：

multiline.pattern: ‘^\[‘

multiline.negate: true

multiline.match: after

 

 

这样，下面的日志就算一个事件了。

 

input

    file

        path => "/var/log/elasticsearch/chuck-clueser.log"

        type => "es-error"

        start_position => "beginning"

        codec => multiline

            pattern => "^\["    #使用正则表式， 以中括号开头的就是一行日志

            negate => true

            what => "previous"

       

   

output

    if [type] == "es-error"

        elasticsearch

            hosts => ["192.168.100.163:9200"]

            index => "es-error-%+YYYY.MM.dd"

        

   

 

2、收集nginx日志

使用codec的json插件将日志的域进行分段，使用key-value的方式，使日志格式更清晰，易于搜索，还可以降低cpu的负载 

2.1 更改nginx的配置文件的日志格式，使用json

[root@linux-node1 ~]# vim /etc/nginx/nginx.conf   #添加日志格式，把自带的格式注释掉

17 http

 18     #log_format  main  ‘$remote_addr - $remote_user [$time_local] "$request"     ‘

 19     #                  ‘$status $body_bytes_sent "$http_referer" ‘

 20     #                  ‘"$http_user_agent" "$http_x_forwarded_for"‘;

 21     #access_log  /var/log/nginx/access.log  main;

 22     log_format json ‘ "@timestamp": "$time_local", ‘

 23                          ‘"@fields": ‘

 24                          ‘"remote_addr": "$remote_addr", ‘

 25                          ‘"remote_user": "$remote_user", ‘

 26                          ‘"body_bytes_sent": "$body_bytes_sent", ‘

 27                          ‘"request_time": "$request_time", ‘

 28                          ‘"status": "$status", ‘

 29                          ‘"request": "$request", ‘

 30                          ‘"request_method": "$request_method", ‘

 31                          ‘"http_referrer": "$http_referer", ‘

 32                          ‘"body_bytes_sent":"$body_bytes_sent", ‘

 33                          ‘"http_x_forwarded_for": "$http_x_forwarded_for", ‘

 34                          ‘"http_user_agent": "$http_user_agent" ‘;

 35     access_log /var/log/nginx/access_json.log json;

[root@linux-node1 ~]# nginx -t  #检查配置文件

[root@linux-node1 ~]# systemctl start nginx

日志格式如下

 

 

2.2 使用logstash将nginx访问日志收集起来

[root@linux-node1 ~]# cat log_nginx.conf 4、

input

    file

        path => "/var/log/nginx/access_json.log"

        codec => "json"

        start_position => "beginning"

        type => "nginx-log"

   

output

    elasticsearch

        hosts => ["http://192.168.100.163:9200"]

        index => "nginx-%+YYY.MM.dd"

       

[root@linux-node1 ~]# /usr/local/logstash/bin/logstash -f log_nginx.conf

 

 

3、收集系统syslog日志

[root@linux-node1 ~]# vim syslog.conf

input

    syslog

        type => "system-syslog"

        #绑定个ip，监听个514端口，启动后，别的机器可以通过网络把日志发过来

        host => "192.168.100.161"

        port => "514"

   

output

    elasticsearch

        hosts => ["192.168.100.161:9200"]

        index => "system-syslog-%+YYYY.MM.dd"

       

   

 

[root@linux-node1 ~]# /usr/local/logstash/bin/logstash -f syslog.conf

 

修改服务器的syslog配置文件，把日志信息发送到514端口上

[root@linux-node2 ~]# vim /etc/rsyslog.conf

90 *.* @@192.168.100.161:514

[root@linux-node2 ~]# systemctl restart rsyslog

 

 

 

4、使用fliter的grok模块收集mysql日志

filter插件有很多，在这里就学习grok插件，使用正则匹配日志里的域来拆分。在实际生产中，apache日志不支持jason，就只能使用grok插件匹配；mysql慢查询日志也是无法拆分，只能使用grok正则表达式匹配拆分。

 

在如下链接，github上有很多写好的grok模板，可以直接引用 

https://github.com/logstash-plugins/logstash-patterns-core/blob/master/patterns/grok-patterns

 

 

在装好的logstash中也会有grok匹配规则，直接可以引用，路径如下

[root@linux-node1 patterns]# pwd

/usr/local/logstash/vendor/bundle/jruby/1.9/gems/logstash-patterns-core-4.1.0/patterns

 

 

 

4.1日志文件

[root@linux-node1 ~]# cat slow.log

# Time: 160108 15:46:14

# User@Host: dev_select_user[dev_select_user] @  [192.168.97.86]  Id: 714519

# Query_time: 1.638396  Lock_time: 0.000163 Rows_sent: 40  Rows_examined: 939155

SET timestamp=1452239174;

SELECT DATE(create_time) as day,HOUR(create_time) as h,round(avg(low_price),2) as low_price

    FROM t_actual_ad_num_log WHERE create_time>=‘2016-01-07‘ and ad_num<=10

    GROUP BY DATE(create_time),HOUR(create_time);

 

4.2编写slow.conf

[root@linux-node1 ~]# cat mysql-slow.conf

input

   file

     path => "/root/slow.log"

     type => "mysql-slow-log"

     start_position => "beginning"

     codec => multiline

        pattern => "^# User@Host:"

        negate => true

        what => "previous"

   

 

filter

      # drop sleep events

    grok

        match => "message" =>"SELECT SLEEP"

        add_tag => [ "sleep_drop" ]

        tag_on_failure => [] # prevent default _grokparsefailure tag on real records

     

     if "sleep_drop" in [tags]

        drop

    

     grok

        match => [ "message", "(?m)^# User@Host: %USER:user\[[^\]]+\] @ (?:(?<clienthost>\S*) )?\[(?:%IP:clientip)?\]\s+Id: %NUMBER:row_id:int\s*# Query_time: %NUMBER:query_time:float\s+Lock_time: %NUMBER:lock_time:float\s+Rows_sent: %NUMBER:rows_sent:int\s+Rows_examined: %NUMBER:rows_examined:int\s*(?:use %DATA:database;\s*)?SET timestamp=%NUMBER:timestamp;\s*(?<query>(?<action>\w+)\s+.*)\n#\s*" ]

     

      date

        match => [ "timestamp", "UNIX" ]

        remove_field => [ "timestamp" ]

     

output

     stdout

     codec => "rubydebug"

  

 

 

执行该配置文件，查看grok正则匹配结果