Home > Article > Backend Development > Two methods for Nginx service log analysis (shell+python)
Two methods for Nginx service log analysis (shell+python)
- 高洛峰Original
- 2017-03-24 15:19:362825browse
python脚本
log_format main ‘$remote_addr – $remote_user [$time_iso8601] “$request” ‘
‘$status $body_bytes_sent “$http_referer” ‘
‘”$http_user_agent” “$http_x_forwarded_for” ‘
‘ “$upstream_addr” “$upstream_status” “$request_time"`;
cat website.access.log| awk ‘{print $(NF)}’ | awk -F “\”” ‘{print $2′}>a.txt
paste -d ” ” website.access.log a.txt > b.txt
cat b.txt |awk ‘($NF>1){print $6$7 ” ” $NF}’>c.txt
linux下使用awk,wc,sort,uniq,grep对nginx日志进行分析和统计
b). 字段含义（如下说明）
column1：ip_address
column2：log_time
column3：request
column4：status_code
column5：send_bytes
column6：referer
需求一：统计总记录数，总成功数，各种失败数：404，403,500
cat data.log|awk -F '\t' '{if($4 > 0) print $4}'|wc -l|
awk '{print "Total Items:"$1}'
2. 提取成功、各种失败总数
cat data.log|awk -F '\t' '{if($4>0 && $4==200) print $4}'|wc -l
需求二：各种错误中，哪类URL出现的次数最多，要求剔除重复项，并倒叙给出结果
cat data.log|awk -F '\t' '{if($4>0 && $4==500) print $3}'|awk '{print $2}'|sort|uniq -c|sort -k1 nr
需求三：要统计URL中文件名出现的次数，结果中要包含Code 和 Referer。但是 URL和 Referer中都包含 / 字符，对于过滤有干扰，尝试去解决。
cat data.log|awk '{print $5,$7,$9}'|grep 200|
sed 's#.*/\(.*\)#\1#'|sort -k1|uniq -c
wc -l access.log |awk '{print $1}' 总请求数
awk '{print $1}' access.log|sort |uniq |wc -l 独立IP数
awk -F'[ []' '{print $5}' access.log|sort|uniq -c|sort -rn|head -5 每秒客户端请求数 TOP5
awk '{print $1}' access.log|sort |uniq -c | sort -rn |head -5 访问最频繁IP Top5
awk '{print $7}' access.log|sort |uniq -c | sort -rn |head -5 访问最频繁的URL TOP5
awk '{if ($12 > 10){print $7}}' access.log|sort|uniq -c|sort -rn |head -5
响应大于10秒的URL TOP5
awk '{if ($13 != 200){print $13}}' access.log|sort|uniq -c|sort -rn|head -5
分析请求数大于50000的源IP的行为
awk '{print $1}' access.log|sort |uniq -c |sort -rn|awk '{if ($1 > 50000){print $2}}' > tmp.txt
for i in $(cat tmp.txt)
do
echo $i >> analysis.txt
echo "访问行为统计" >> analysis.txt
grep $i access.log|awk '{print $6}' |sort |uniq -c | sort -rn |head -5 >> analysis.txt
echo "访问接口统计" >> analysis.txt
grep $i access.log|awk '{print $7}' |sort |uniq -c | sort -rn |head -5 >> analysis.txt
echo -e "\n" >> /root/analysis/$Ydate.txt
done
如果源IP来自代理服务器，应将第一条命令过滤地址改为$http_x_forwarded_for地址
awk '{print $NF}' access.log|sort |uniq -c |sort -rn|awk '{if ($1 > 50000){print $2}}' > tmp.txt
5.性能指标
并发连接数
客户端向服务器发起请求，并建立了TCP连接。每秒钟服务器链接的总TCP数量，就是并发连接数
PV（page view） UV（unique visitor） 独立IP
6.故障
1.Nginx Connection 不夠用 的參數調整
2.nginx+php-fpm出现502
3.线上nginx的一次“no live upstreams while connecting to upstream ”分析
4.nginx proxy_pass末端神奇的斜线
5.nginx+tomcat使用apache的FtpClient上传图片时由于多线程问题导致的文件大小为0的问题
案例一
ip - - [23/Mar/2017:00:17:49 +0800] "GET / HTTP/1.1" 302 0 "-" "PycURL/7.19.7"
log_format access '$HTTP_X_REAL_IP - $remote_user [$time_local] "$request"'
'$status $body_bytes_sent "$http_referer" '
'"$http_user_agent" $HTTP_X_Forwarded_For';
192.168.21.1 - - [27/Jan/2014:11:28:53 +0800] "GET /2.php HTTP/1.1" 200 133 "-" "Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/33.0.1707.0 Safari/537.36" "-"192.168.21.128 200 127.0.0.1:9000 0.119 0.119
#log_format main '$remote_addr - $remote_user [$time_local] "$request" '
# '$status $body_bytes_sent "$http_referer" '
# '"$http_user_agent" "$http_x_forwarded_for"';
$http_host：用户在浏览器中输入的URL（IP或着域名）地址 192.168.21.128
$upstream_status： upstream状态 200
$upstream_addr： 后端upstream地址及端口 127.0.0.1:9000
$request_time： 页面访问总时间 0.119
$upstream_response_time：页面访问中upstream响应时间 0.119
$10 $body_bytes_sent
$1 $remote_addr
$7 $request
$11 $http_referer
$9 $status
$6 http_user_agent
1、总访问量
2、总带宽
3、独立访客量
4、访问IP统计
5、访问url统计
6、来源统计
7、404统计
8、搜索引擎访问统计(谷歌，百度)
9、搜索引擎来源统计(谷歌，百度)
#!/bin/bash
log_path=/home/www.centos.bz/log/access.log.1
domain="centos.bz"
email="log@centos.bz"
maketime=`date +%Y-%m-%d" "%H":"%M`
logdate=`date -d "yesterday" +%Y-%m-%d`
total_visit=`wc -l ${log_path} | awk '{print $1}'`
total_bandwidth=`awk -v total=0 '{total+=$10}END{print total/1024/1024}' ${log_path}`
total_unique=`awk '{ip[$1]++}END{print asort(ip)}' ${log_path}`
ip_pv=`awk '{ip[$1]++}END{for (k in ip){print ip[k],k}}' ${log_path} | sort -rn | head -20`
url_num=`awk '{url[$7]++}END{for (k in url){print url[k],k}}' ${log_path} | sort -rn | head -20`
referer=`awk -v domain=$domain '$11 !~
/http:\/\/[^/]*'"$domain"'/{url[$11]++}END{for (k in url){print
url[k],k}}' ${log_path} | sort -rn | head -20`
notfound=`awk '$9 == 404 {url[$7]++}END{for (k in url){print url[k],k}}' ${log_path} | sort -rn | head -20`
spider=`awk -F'"' '$6 ~ /Baiduspider/ {spider["baiduspider"]++} $6 ~
/Googlebot/ {spider["googlebot"]++}END{for (k in spider){print
k,spider[k]}}' ${log_path}`
search=`awk -F'"' '$4 ~ /http:\/\/www\.baidu\.com/
{search["baidu_search"]++} $4 ~ /http:\/\/www\.google\.com/
{search["google_search"]++}END{for (k in search){print k,search[k]}}'
${log_path}`
#echo -e "概况\n报告生成时间：${maketime}\n总访问量:${total_visit}\n总带宽:${total_bandwidth}M\n独
立访客:${total_unique}\n\n访问IP统计\n${ip_pv}\n\n访问url统计\n${url_num}\n\n来源页面统计
\n${referer}\n\n404统计\n${notfound}\n\n蜘蛛统计\n${spider}\n\n搜索引擎来源统计
\n${search}" | mail -s "$domain $logdate log statistics" ${email}案例二
# tar zxvf pymongo-1.11.tar.gz
# cd pymongo-1.11
# python setup.py install
python连接mongodb样例
$ cat conn_mongodb.py
#!/usr/bin/python
import pymongo
import random
conn = pymongo.Connection("127.0.0.1",27017)
db = conn.tage #连接库
db.authenticate("tage","123")
#用户认证
db.user.drop()
#删除集合user
db.user.save({'id':1,'name':'kaka','sex':'male'})
#插入一个数据
for id in range(2,10):
name = random.choice(['steve','koby','owen','tody','rony'])
sex = random.choice(['male','female'])
db.user.insert({'id':id,'name':name,'sex':sex})
#通过循环插入一组数据
content = db.user.find()
#打印所有数据
for i in content:
print i
编写python脚本
#encoding=utf8
import re
zuidaima_nginx_log_path="/usr/local/nginx/logs/www.zuidaima.com.access.log"
pattern = re.compile(r'^\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}')
def stat_ip_views(log_path):
ret={}
f = open(log_path, "r")
for line in f:
match = pattern.match(line)
if match:
ip=match.group(0)
if ip in ret:
views=ret[ip]
else:
views=0
views=views+1
ret[ip]=views
return ret
def run():
ip_views=stat_ip_views(zuidaima_nginx_log_path)
max_ip_view={}
for ip in ip_views:
views=ip_views[ip]
if len(max_ip_view)==0:
max_ip_view[ip]=views
else:
_ip=max_ip_view.keys()[0]
_views=max_ip_view[_ip]
if views>_views:
max_ip_view[ip]=views
max_ip_view.pop(_ip)
print "ip:", ip, ",views:", views
#总共有多少ip
print "total:", len(ip_views)
#最大访问的ip
print "max_ip_view:", max_ip_view
run()The above is the detailed content of Two methods for Nginx service log analysis (shell+python). For more information, please follow other related articles on the PHP Chinese website!