您的位置：首页 > 其它

ELK+Kafka 企业日志收集平台(二)这是原版

2017-09-06 10:48 537 查看

上篇博文主要总结了一下elk、基于kafka的zookeeper集群搭建，以及系统日志通过zookeeper集群达到我们集群的整个过程。下面我们接着下面这个未完成的几个主题

4.Kibana部署;

5.Nginx负载均衡Kibana请求;

6.案例：nginx日志收集以及MySQL慢日志收集;

7.Kibana报表基本使用;

Kibana的部署;

Kibana的作用，想必大家都知道了就是一个展示工具，报表内容非常的丰富；

下面我们在两台es上面搭建两套kibana

1.获取kibana软件包

[root@es1 ~]# wget https://download.elastic.co/kibana/kibana/kibana-4.1.2-linux-x64.tar.gz [root@es1 ~]# tar -xf kibana-4.2.0-linux-x64.tar.gz -C /usr/local/

2.修改配置文件

[root@es1 ~]# cd /usr/local/
[root@es1 local]# ln -sv kibana-4.1.2-linux-x64 kibana
`kibana' -> `kibana-4.2.0-linux-x64'
[root@es1 local]# cd kibana

[root@es1 kibana]# vim config/kibana.yml
server.port: 5601      #默认端口可以修改的
server.host: "0.0.0.0" #kibana监听的ip
elasticsearch.url: "http://localhost:9200" #由于es在本地主机上面，所以这个选项打开注释即可

3.提供kibana服务管理脚本，我这里写了个相对简单的脚本

[root@es1 config]# cat /etc/init.d/kibana
#!/bin/bash
#chkconfig: 2345 55 24
#description: kibana service manager

KIBBIN='/usr/local/kibana/bin/kibana'
LOCK='/usr/local/kibana/locks'

START() {
if [ -f $LOCK ];then
echo -e "kibana is already \033[32mrunning\033[0m, do nothing."
else
echo -e "Start kibana service.\033[32mdone\033[m"
cd  /usr/local/kibana/bin
nohup ./kibana & >/dev/null
touch $LOCK
fi
}

STOP() {
if [ ! -f $LOCK ];then
echo -e "kibana is already stop, do nothing."
else
echo -e "Stop kibana serivce \033[32mdone\033[m"
rm -rf $LOCK
ps -ef | grep kibana | grep -v "grep" | awk '{print $2}' | xargs kill -s 9 >/dev/null
fi
}

STATUS() {
Port=$(netstat -tunl | grep ":5602")
if [ "$Port" != "" ] && [ -f $LOCK ];then
echo -e "kibana is: \033[32mrunning\033[0m..."
else
echo -e "kibana is: \033[31mstopped\033[0m..."
fi
}

case "$1" in
start)
START
;;
stop)
STOP
;;
status)
STATUS
;;
restart)
STOP
sleep 2
START
;;
*)
echo "Usage: /etc/init.d/kibana (|start|stop|status|restart)"
;;
esac

4.启动kibana服务

[root@es1 config]# chkconfig --add kibana
[root@es1 config]# service kibana start
Start kibana service.done
[root@es1 config]#

5.服务检查

[root@es1 config]# ss -tunl | grep "5601"
tcp    LISTEN     0      511                    *:5601                  *:*
[root@es1 config]#

ok，此时我直接访问es1这台主机的5601端口

ok，能成功的访问5601端口，那我把es1这台的配置放到es2上面去然后启动，效果跟访问es1一样

Nginx负载均衡kibana的请求

1.在nginx-proxy上面yum安装nginx

yum install -y nignx

2.编写配置文件es.conf

[root@saltstack-node1 conf.d]# pwd
/etc/nginx/conf.d
[root@saltstack-node1 conf.d]# cat es.conf
upstream es {
server 192.168.2.18:5601 max_fails=3 fail_timeout=30s;
server 192.168.2.19:5601 max_fails=3 fail_timeout=30s;
}

server {
listen       80;
server_name  localhost;

location / {
proxy_pass http://es/; index index.html index.htm;
#auth
auth_basic "ELK Private";
auth_basic_user_file /etc/nginx/.htpasswd;
}

}

3.创建认证

[root@saltstack-node1 conf.d]# htpasswd -cm /etc/nginx/.htpasswd elk
New password:
Re-type new password:
Adding password for user elk-user
[root@saltstack-node1 conf.d]# /etc/init.d/nginx restart
Stopping nginx:                                            [  OK  ]
Starting nginx:                                            [  OK  ]
[root@saltstack-node1 conf.d]#

4.直接输入认证用户及密码就可访问啦http://192.168.2.21/

Nginx及MySQL慢日志收集

首先我们在webserver1上面都分别安装了nginx 及mysql.

1.为了方便nginx日志的统计搜索，这里设置nginx访问日志格式为json

(1)修改nginx主配置文件

说明：如果想实现日志的报表展示，最好将业务日志直接以json格式输出，这样可以极大减轻cpu负载，也省得运维需要写负载的filter过滤正则。

[root@webserver1 nginx]# vim nginx.conf
log_format json '{"@timestamp":"$time_iso8601",'
'"@version":"1",'
'"client":"$remote_addr",'
'"url":"$uri",'
'"status":"$status",'
'"domain":"$host",'
'"host":"$server_addr",'
'"size":$body_bytes_sent,'
'"responsetime":$request_time,'
'"referer": "$http_referer",'
'"ua": "$http_user_agent"'
'}';
access_log  /var/log/access_json.log  json;

(2)收集nginx日志和MySQL日志到消息队列中；这个文件我们是定义在客户端，即生产服务器上面的Logstash文件哦.

注意：这里刚搭建完毕，没有什么数据，为了展示效果，我这里导入了线上的nginx和MySQL慢日志

input {
file {             #从nginx日志读入
type => "nginx-access"
path => "/var/log/nginx/access.log"
start_position => "beginning"
codec => "json"  #这里指定 codec格式为json
}
file {  #从MySQL慢日志读入
type => "slow-mysql"
path => "/var/log/mysql/slow-mysql.log"
start_position => "beginning"
codec => multiline {         #这里用到了logstash的插件功能，将本来属于一行的多行日志条目整合在一起，让他属于一条
pattern => "^# User@Host"  #用到了正则去匹配
negate => true
what => "previous"
}
}
}

output {
#  stdout { codec=> rubydebug }
if [type] == "nginx-access" {    #通过判断input中定义的type，来让它在kafka集群中生成的主题名称
kafka {                        #输出到kafka集群
bootstrap_servers => "192.168.2.22:9092,192.168.2.23:9092,192.168.2.24:9092"  #生产者们
topic_id => "nginx-access"   #主题名称
compression_type => "snappy" #压缩类型
}
}
if [type] == "slow-mysql" {
kafka {
bootstrap_servers => "192.168.2.22:9092,192.168.2.23:9092,192.168.2.24:9092"
topic_id => "slow-mysql"
compression_type => "snappy"
}
}
}

(3)Logstash 从kafka集群中读取日志存储到es中，这里的定义logstash文件是在三台kafka服务器上面的哦，并且要保持一致，你可以在一台上面修改测试好之后，拷贝至另外两台即可。

input {
kafka {
zk_connect => "192.168.2.22:2181,192.168.2.23:2181,192.168.2.24:2181"
type => "nginx-access"
topic_id => "nginx-access"
codec => plain
reset_beginning => false
consumer_threads => 5
decorate_events => true
}
kafka {
zk_connect => "192.168.2.22:2181,192.168.2.23:2181,192.168.2.24:2181"
type => "slow-mysql"
topic_id => "slow-mysql"
codec => plain
reset_beginning => false
consumer_threads => 5
decorate_events => true
}
}

output {
#  stdout { codec=> rubydebug }
if [type] == "nginx-access" {
elasticsearch {
hosts => ["192.168.2.18:9200","192.168.2.19:9200"]
index => "nginx-access-%{+YYYY-MM}"
}
}
if [type] == "slow-mysql" {
elasticsearch {
hosts => ["192.168.2.18:9200","192.168.2.19:9200"]
index => "slow-mysql-%{+YYYY-MM}"
}
}
}