部署 elk 日志系统 elasticsearch、logstash、 kibana
安装elk安装Java 1.8环境
解压源码安装包：
tar xf jdk-8u121-linux-x64.tar.gz
ll
mkdir /work/opt -p
mvjdk1.8.0_121/work/opt/jdk
ll /work/opt/jdk/
chown -R root.root/work/opt
vim /etc/profile : //添加
export JAVA_HOME=/work/opt/jdk
export JAVA_BIN=$JAVA_HOME/bin
export JRE_HOME=${JAVA_HOME}/jre
export CLASSPATH=$JAVA_HOME/jre/lib/ext:$JAVA_HOME/lib/tools.jar
export PATH=$JAVA_BIN:$PATH
安装 elasticsearch-5.3.0
tar xf elasticsearch-5.3.0.tar.gz
mv elasticsearch-5.3.0 /work/opt/elasticsearch
ubuntu@ip-172-31-1-79:/work/opt/elasticsearch/config$ egrep -v '#|^$' elasticsearch.yml
cluster.name: lvnian-elk
node.name: lvnian-elk-node1
path.data: /data#由于是普通用户启动elasticsearch，所以这个目录的属主需要改为普通用户
path.logs: /work/opt/elasticsearch/logs #由于是普通用户启动elasticsearch，所以这个目录的属主需要改为普通用户
bootstrap.memory_lock: false
network.host: 0.0.0.0
http.port: 9200
http.cors.enabled: true
http.cors.allow-origin: "*"
####缓存
index.cache.field.expire: 10m
index.cache.field.max_size: 500000
index.cache.field.type: soft
ubuntu@ip-172-31-1-79:/work/opt/elasticsearch/config$
nohup /work/opt/elasticsearch/bin/elasticsearch >> /tmp/elasticsearch.log& ##ES 5.1.1 安装 head：
（5.1.1版本的 elasticsearch 没有提供直接插件安装方法，但在该github上该插件作者给出了方法）
下载二进制源码包：
wget https://nodejs.org/dist/v4.2.2/node-v6.2.0-linux-x64.tar.gz
解压：
tar xf node-v6.2.0-linux-x64.tar.gz -C /work/opt/
设置环境变量：
vim /etc/profile：
export NODE_HOME=/work/opt/node/
export PATH=$PATH:$NODE_HOME/bin
root@ip-172-31-1-79:/work/source# node --version
v6.10.1
npm config set registry https://registry.npm.taobao.org //设置代理镜像源，加速下载
cd /home/stt/node-v4.2.2-linux-x64/lib/node_modules
npm install grunt //显示的2条warn可以忽略 测试grunt是否生效：
$ grunt -version
grunt-cli v1.2.0
grunt v1.0.1 安装head插件：
下载：git clone git://github.com/mobz/elasticsearch-head.git
cd /home/stt/elasticsearch-head
npm install (提示：如果遇到网络的瓶颈，将预先下载的源码包放在对应的位置效果一样，目录为/tmp/phantomjs/，确定位置后可自行创建并上传包)
修改 elasticsearch-head/ _site/app.js
// 把localhost改为ip
找到：
this.base_uri = this.config.base_uri || this.prefs.get("app-base_uri") || "http://localhost:9200";
改为：
this.base_uri = this.config.base_uri || this.prefs.get("app-base_uri") || "http://192.168.8.116:9200";
修改 elasticsearch-head/Gruntfile.js
connect: {
server: {
options: {
hostname: "0.0.0.0", //增加这个配置
port: 9100,
base: '.',
keepalive: true
}
}
} 启动服务：（后台运行）
grunt server & //需要在 /home/stt/elasticsearch-head 下执行，因为我的 grunt 没有进行全局的安装
##安装Logstash(这个软件在你需要读取的日志服务上安装，用logstash读取你的日志，上传给elasticsearch的):
也需要安装java1.8环境
tar xf logstash-5.3.0.tar.gz
mkdir /work/opt
mv logstash-5.3.0 /work/opt/
cd /work/opt/ vim /work/opt/logstash-5.3.0/conf/central.conf #(处理基于 FILE 方式输入的日志信息，这里是简单的举个例子，日后继续学习补充)
input {
file {
path => "/tmp/*.log"
}
}
output {
elasticsearch {
hosts => "192.168.8.116:9200"
index => "nginx-access"
}
stdout {
codec => rubydebug
}
}
##安装Kibana：
解压源码包：
tar zxf kibana-5.1.1-linux-x86_64.tar.gz -C /home/stt/server/
vim config/kibana.yml //修改
server.port: 5601 //打开注释而已，不用可以去效果，请使用默认端口
server.host: "0.0.0.0" //打开监听地址，让别的机器也能访问这个 kibana
elasticsearch.url: "http://127.0.0.1:9200" //这个url要根据实质情况，添加访问 elasticsearch 的url
启动服务： （后台运行）
/home/stt/server/kibana-5.1.1-linux-x86_64/bin/kibana &
安装nginx 反向代理，
apt-gei install nginx
nginx放心代理配置文件如下：
##文件名kibana.conf
upstream backend {
server 172.31.6.155:5601;
}
server {
listen 80;
server_namekibana.lvnian.co;
access_log /tmp/kibana-access.log;
error_log /tmp/kibana-error.log;
location / {
#设置主机头和客户端真实地址，以便服务器获取客户端真实IP
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
#禁用缓存
proxy_buffering off;
#反向代理的地址
proxy_pass http://backend;
}
}
logstash 读取nginx的访问日志以及error日志，上传到logstash的配置文件文件。
用下面命令运行即可
nohup /work/opt/logstash-5.3.0/bin/logstash -f /work/opt/logstash-5.3.0/conf/elk-nginx-log.conf &
文件名： elk-nginx-log.conf
input {
file {
path => "/data/logs/nginx/*.log"
start_position => beginning
}
}
filter {
if =~ "access"{
mutate { replace => { type => "nginx_access" } }
ruby {
init => "@kname = ['http_x_forwarded_for','http_host','time_local','request','status','body_bytes_sent','request_body','content_length','http_referer','http_user_agent','http_cookie','remote_addr','remote_port','hostname','upstream_addr','upstream_response_time','request_time']"
code => "
new_event = LogStash::Event.new(Hash[@kname.zip(event.get('message').split(' || '))])
new_event.remove('@timestamp')
event.append(new_event)
"
}
if {
ruby {
init => "@kname = ['method','uri','verb']"
code => "
new_event = LogStash::Event.new(Hash[@kname.zip(event.get('request').split(' '))])
new_event.remove('@timestamp')
event.append(new_event)
"
}
if {
ruby {
init => "@kname = ['url_path','url_args']"
code => "
new_event = LogStash::Event.new(Hash[@kname.zip(event.get('uri').split('?'))])
new_event.remove('@timestamp')
event.append(new_event)
"
}
}
}
mutate {
convert => [
"body_bytes_sent" , "integer",
"content_length", "integer",
"upstream_response_time", "float",
"request_time", "float",
"http_x_forwarded_for", "string",
"http_host", "string"
]
remove_field => [ "message","uri","request","path","verb" ]
}
if == '-' or '.' notin {
mutate { replace => { http_x_forwarded_for => "%{remote_addr}" } } ##http_x_forwarded_for 字段为空，则把remote_addr字段的内容赋值给它
}
if !~ "^127\.|^192\.168\.|^172\.1\.|^172\.2\.|^172\.3\.|^10\." {
geoip {
source => "remote_addr" #设置解析IP地址的字段
target => "geoip" #将geoip数据保存到一个字段内
database => "/work/opt/logstash/conf/GeoLite2-City.mmdb" #IP地址数据库
add_field => [ "", "%{}" ]
add_field => [ "", "%{}"]
}
mutate {
convert => [ "", "float"]
}
}
date {
match => [ "time_local", "dd/MMM/yyyy:hh:mm:ss Z" ]
locale => "zh"
}
}
else if =~ "error" {
mutate { replace => { type => "nginx_error" } }
grok {
match => { "message" => "(?\d\d\d\d/\d\d/\d\d \d\d:\d\d:\d\d) \[(?\w+)\] \S+: \*\d+ (?[^,]+), (?.*)$" }
}
mutate {
rename => [ "host", "fromhost" ]
gsub => [ "errmsg", "too large body: \d+ bytes", "too large body" ]
}
if
{
ruby {
code => "
new_event = LogStash::Event.new(Hash)
new_event.remove('@timestamp')
event.append(new_event)
"
}
}
grok {
# match => { "request" => '"%{WORD:verb} %{URIPATH:urlpath}(?:\?%{NGX_URIPARAM:urlparam})?(?: HTTP/%{NUMBER:httpversion})"' }
match => { "request" => '"%{WORD:verb} %{URIPATH:urlpath}?(?: HTTP/%{NUMBER:httpversion})"' }
patterns_dir => ["/etc/logstash/patterns"]
# remove_field => [ "message", "errinfo", "request" ]
}
}
else {
mutate { replace => { type => "random_logs" } }
}
#################时区问题解决
ruby {
code => "event.set('timestamp', event.get('@timestamp').time.localtime + 8*60*60)"
}
ruby {
code => "event.set('@timestamp',event.get('timestamp'))"
}
mutate {
remove_field => ["timestamp"]
}
###############
}
output {
elasticsearch {
hosts => "10.19.104.161:9200"
#index => "logstash-nginx"
index => "logstash-%{type}-%{+YYYY.MM.dd}"
document_type => "%{type}"
flush_size => 20000
idle_flush_time => 10
sniffing => true
template_overwrite => true
}
stdout {
codec => rubydebug
}
} 使用上面的logstash配置文件，需要注意更改nginx日志格式，nginx日志格式应该改为如下：
log_format elk"$http_x_forwarded_for | $time_local | $request | $status | $body_bytes_sent | "
"$request_body | $content_length | $http_referer | $http_user_agent| "
"$http_cookie | $remote_addr | $hostname | $upstream_addr | $upstream_response_time | $request_time|$gzip_ratio";
access_log /data/log/nginx/access.log elk;
error_log/data/log/nginx/error.log; 监控5分钟内xxx 状态码 出现 阀值n 次进行的微信报警
#!/usr/bin/env python
# -*- coding:utf-8 -*-
#Author: gaogd
import weixin_alter asweixin
from elasticsearch import Elasticsearch
import json,time ,datetime,os
class Checkcode(object):
def __init__(self,statuscode,num=30):
self.host='172.31.1.79'
self.port=9200
self.statuscode=statuscode
self.es = Elasticsearch([{'host': self.host, 'port': self.port}])
#now_day = time.strftime('%Y.%m.%d', time.localtime(time.time()))
now_day = (datetime.datetime.now() + datetime.timedelta(hours=-8)).strftime("%Y.%m.%d")
## 8 个小时差是为了解决logstash 导入日志到es中，索引回更加0时区来创建索引的，导致索引日期相差8个小时
self.index = 'logstash-nginx_access-%s' % now_day
self.file='./alter%s.log'%self.statuscode
self.num=int(num)
def Getdata(self):
body = {
"query": {
"bool": {
"filter": [
{"term":
{"status": "%s"%self.statuscode
}
},
{
"range": {
"@timestamp": {
"gt": "now-5m",
"lt": "now"
},
},
},
]
}
},
"aggs": {
"api_status_total": {"terms": {"field": "url_path.keyword"}}
},
"size": 0
}
res = self.es.search(index=self.index, body=body)
num = res['hits']['total']
api_list=res['aggregations']['api_status_total']['buckets']
dict_api = {}
for l in api_list:
if type(l) is dict:
dict_api]=l['doc_count']
dictlist = sorted(dict_api.iteritems(), key=lambda d: d, reverse=True)
content=''
for n,v in enumerate(dictlist):
if n > 9:
break
content = content + v.encode('utf-8') + " ： " + str(v) + "\n"
print 'code:', code, 'time', num
print content
return num,content
def JudgeAlter(self,num,content):
if not os.path.isfile(self.file):
with open(self.file, 'w') as f:
f.write('0')
if int(num) > self.num:
with open(self.file, 'r') as f:
time = f.read()
if len(time) == 0 or int(time) == 0 or int(time) >= 15:
content = u"""
%s告警!!!
最近5分钟内前端日志出现%s错误\n 5 分钟内有： %s次
%s"""% (self.statuscode,self.statuscode,num,content)
print content
weixin.WeixinSend(str(content))
with open(self.file, 'w+') as f:
f.write('1')
elif 0 < int(time)0:
content = u"""
%s 告警恢复！！！
最近5分钟内日志出现%s 错误少于%s次
现在错误次数为： %s次
%s"""% (self.statuscode,self.statuscode,self.num,num,content)
print content
weixin.WeixinSend(str(content))
with open(self.file, 'w+') as f:
f.write('0')
if __name__ == "__main__":
print datetime.datetime.now()
code='500'
num=30
try:
import sys
code=sys.argv
num=sys.argv
exceptException as err:
print str(err)
print "Usage: python %s codenum"%sys.argv
exit()
obj=Checkcode(code,num)
num,dict_api=obj.Getdata()
obj.JudgeAlter(num,dict_api)
### python /CheckStatusCode.py 500 20 logstash分析日志，导入到es中的时区问题解决
input { stdin {} }
output { stdout { codec => rubydebug } }
filter {
date {
match => ["message","UNIX_MS"]
target => "@timestamp"
}
ruby {
code => "event.set('timestamp', event.get('@timestamp').time.localtime + 8*60*60)"
}
ruby {
code => "event.set('@timestamp',event.get('timestamp'))"
}
mutate {
remove_field => ["timestamp"]
}
}
上面这种方法，实际上是把时间多加8小时，时区并没有发生变化，这样做有一个坏处，就是你的kibana默认
是CST时区了，这个时候展示出来的数据就会在8个小时之后的地方了。有问题，不建议使用。
市区问题其实不用修改，kibana能够很好识别。 ## 参考 elk 框架
flume| ===èkafka==èlogstash收集==èes 存储==è kibana展示
页:
[1]