1. 概述
1.1 相关地址
1.2 组件分工
- filebeat：部署在每台应用服务器、数据库、中间件中，负责日志抓取与日志聚合
- 日志聚合：把多行日志合并成一条，例如exception的堆栈信息等
- logstash：通过各种filter结构化日志信息，并把字段transform成对应的类型
- elasticsearch：负责存储和查询日志信息
- kibana：通过ui展示日志信息、还能生成饼图、柱状图等
2. docker安装部署
step1：修改mmap计数大于等于262144的限制
#在/etc/sysctl.conf文件最后添加一行 vm.max_map_count=655360 #并执行命令 sysctl -p
step2：下载并运行镜像
docker run -p 5601:5601 -p 9200:9200 -p 9300:9300 -p 5044:5044 --name elk -d sebp/elk:651
step3：准备elasticsearch的配置文件
mkdir /opt/elk/elasticsearch/conf #复制elasticsearch的配置出来 docker cp elk:/etc/elasticsearch/elasticsearch.yml /opt/elk/elasticsearch/conf
step4：修改elasticsearch.yml配置
修改cluster.name参数
cluster.name: my-e
在最后新增以下三个参数：
thread_pool.bulk.queue_size: 1000 http.cors.enabled: true http.cors.allow-origin: "*"
step5：准备logstash的配置文件
mkdir /opt/elk/logstash/conf #复制logstash的配置出来 docker cp elk:/etc/logstash/conf.d/. /opt/elk/logstash/conf/
step6：准备logstash的patterns文件
新建一个java的patterns文件
mkdir /opt/elk/logstash/patterns
vim java 内容如下:
# user-center MYAPPNAME ([0-9a-zA-Z_-]*) # RMI TCP Connection(2)-127.0.0.1 MYTHREADNAME ([0-9a-zA-Z._-]|\(|\)|\s)*
就是一个名字叫做java的文件，不需要文件后缀
step7：删除02-beats-input.conf的最后三句，去掉强制认证
vim /opt/elk/logstash/conf/02-beats-input.conf #ssl => true #ssl_certificate => "/pki/tls/certs/logstash.crt" #ssl_key => "/pki/tls/private/logstash.key"
step8：修改10-syslog.conf配置，改为以下内容
- 注意，下面的logstash结构化配置样例都是以本工程的日志格式配置，并不是通用的
filter { if [type] == "syslog" { grok { match => { "message" => "%{SYSLOGTIMESTAMP:syslog_timestamp} %{SYSLOGHOST:syslog_hostname} %{DATA:syslog_program}(?:\[%{POSINT:syslog_pid}\])?: %{GREEDYDATA:syslog_message}" } add_field => [ "received_at", "%{@timestamp}" ] add_field => [ "received_from", "%{host}" ] } syslog_pri { } date { match => [ "syslog_timestamp", "MMM d HH:mm:ss", "MMM dd HH:mm:ss" ] } } if [fields][docType] == "sys-log" { grok { patterns_dir => ["/opt/elk/logstash/patterns"] match => { "message" => "\[%{NOTSPACE:appName}:%{NOTSPACE:serverIp}:%{NOTSPACE:serverPort}\] %{TIMESTAMP_ISO8601:logTime} %{LOGLEVEL:logLevel} %{WORD:pid} \[%{MYAPPNAME:traceId}\] \[%{MYTHREADNAME:threadName}\] %{NOTSPACE:classname} %{GREEDYDATA:message}" } overwrite => ["message"] } date { match => ["logTime","yyyy-MM-dd HH:mm:ss.SSS Z"] } date { match => ["logTime","yyyy-MM-dd HH:mm:ss.SSS"] target => "timestamp" locale => "en" timezone => "+08:00" } mutate { remove_field => "logTime" remove_field => "@version" remove_field => "host" remove_field => "offset" } } if [fields][docType] == "point-log" { grok { patterns_dir => ["/opt/elk/logstash/patterns"] match => { "message" => "%{TIMESTAMP_ISO8601:logTime}\|%{MYAPPNAME:appName}\|%{WORD:resouceid}\|%{MYAPPNAME:type}\|%{GREEDYDATA:object}" } } kv { source => "object" field_split => "&" value_split => "=" } date { match => ["logTime","yyyy-MM-dd HH:mm:ss.SSS Z"] } date { match => ["logTime","yyyy-MM-dd HH:mm:ss.SSS"] target => "timestamp" locale => "en" timezone => "+08:00" } mutate { remove_field => "logTime" remove_field => "@version" remove_field => "host" remove_field => "offset" } } }
step9：修改30-output.conf配置，改为以
output { if [fields][docType] == "sys-log" { elasticsearch { hosts => ["localhost"] manage_template => false index => "sys-log-%{+YYYY.MM.dd}" document_type => "%{[@metadata][type]}" } } if [fields][docType] == "point-log" { elasticsearch { hosts => ["localhost"] manage_template => false index => "point-log-%{+YYYY.MM.dd}" document_type => "%{[@metadata][type]}" routing => "%{type}" } } }
step10：创建运行脚本
vim /opt/elk/start.sh
docker stop elk docker rm elk docker run -p 5601:5601 -p 9200:9200 -p 9300:9300 -p 5044:5044 \ -e LS_HEAP_SIZE="1g" -e ES_JAVA_OPTS="-Xms2g -Xmx2g" \ -v $PWD/elasticsearch/data:/var/lib/elasticsearch \ -v $PWD/elasticsearch/plugins:/opt/elasticsearch/plugins \ -v $PWD/logstash/conf:/etc/logstash/conf.d \ -v $PWD/logstash/patterns:/opt/logstash/patterns \ -v $PWD/elasticsearch/conf/elasticsearch.yml:/etc/elasticsearch/elasticsearch.yml \ -v $PWD/elasticsearch/log:/var/log/elasticsearch \ -v $PWD/logstash/log:/var/log/logstash \ --name elk \ -d sebp/elk:651
step11：运行镜像
sh s
step12：添加索引模板(非必需)
如果是单节点的es需要去掉索引的副本配置，不然会出现unassigned_shards
1.更新已有索引
curl -X PUT "http://192.168.28.130:9200/sys-log-*/_settings" -H 'Content-Type: application/json' -d' { "index" : { "number_of_replicas" : 0 } } ' curl -X PUT "http://192.168.28.130:9200/mysql-slowlog-*/_settings" -H 'Content-Type: application/json' -d' { "index" : { "number_of_replicas" : 0 } }'
2.设置索引模板
系统日志
curl -XPUT http://192.168.28.130:9200/_template/template_sys_log -H 'Content-Type: application/json' -d ' { "index_patterns" : ["sys-log-*"], "order" : 0, "settings" : { "number_of_replicas" : 0 }, "mappings": { "doc": { "properties": { "message": { "type": "text", "fields": { "keyword": { "type": "keyword", "ignore_above": 256 } }, "analyzer": "ik_max_word" }, "pid": { "type": "text" }, "serverPort": { "type": "text" }, "logLevel": { "type": "text" }, "traceId": { "type": "text" } } } } }'
慢sql日志
curl -XPUT http://192.168.28.130:9200/_template/template_sql_slowlog -H 'Content-Type: application/json' -d ' { "index_patterns" : ["mysql-slowlog-*"], "order" : 0, "settings" : { "number_of_replicas" : 0 }, "mappings": { "doc": { "properties": { "query_str": { "type": "text", "fields": { "keyword": { "type": "keyword", "ignore_above": 256 } }, "analyzer": "ik_max_word" } } } } }'
埋点日志
curl -XPUT http://192.168.28.130:9200/_template/template_point_log -H 'Content-Type: application/json' -d ' { "index_patterns" : ["point-log-*"], "order" : 0, "settings" : { "number_of_shards" : 2, "number_of_replicas" : 0 } }'
step13：安装IK分词器
查询数据，都是使用的默认的分词器，分词效果不太理想，会把text的字段分成一个一个汉字，然后搜索的时候也会把搜索的句子进行分词，所以这里就需要更加智能的分词器IK分词器了
1.下载
- 下载地址：https://github.com/medcl/elasticsearch-analysis-ik/releases
- 这里你需要根据你的Es的版本来下载对应版本的IK
2.解压
将文件复制到 es的安装目录/plugin/ik下面即可，完成之后效果如下：
3.重启容器并检查插件是否安装成功
http://192.168.28.130:9200/_cat/plugins
step14：配置样例
链接: https://pan.baidu.com/s/1Qq3ywAbXMMRYyYxBAViMag
提取码: aubp
