nginx日志分析
2017-10-27 本文已影响68人
kekefund
1，日志定义
在nginx.conf中定义的日志格式如下：
http {
...
log_format main '$remote_addr - $remote_user [$time_local] "$request" '
'$status [$request_body] $body_bytes_sent "$http_referer" '
'"$http_user_agent" "$http_x_forwarded_for"';
...
}
日志文件如下：
116.2.52.247 - - [26/Oct/2017:15:04:00 +0000] "POST /api/v1/f1_static/ HTTP/1.1" 200 [{\x22user_id\x22:\x229b999d46dd6149f49\x22}] 323 "http://www.abc.com/ProductPerspective/detail/" "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_12_6) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/61.0.3163.100 Safari/537.36" "-"
116.2.52.247 - - [26/Oct/2017:15:04:00 +0000] "OPTIONS /api/v1/fund_info/ HTTP/1.1" 200 [-] 31 "http://www.abc.com/ProductPerspective/detail/" "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_12_6) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/61.0.3163.100 Safari/537.36" "-"
2，日志分割
nginx没有命令直接将日志按天分割，我们写了一个shell脚本，每日0点定时执行。
#nginx.log.sh
#nginx日志切割脚本
#!/bin/bash
#设置日志文件存放目录
logs_path="/mydata/nginx/logs/"
#重命名日志文件
mv ${logs_path}access-web.log ${logs_path}access-web-$(date -d "yesterday" +"%Y%m%d").log
mv ${logs_path}access-api.log ${logs_path}access-api-$(date -d "yesterday" +"%Y%m%d").log
cron：
0 0 * * * /mydata/nginx/nginx.log.sh
3，日志搜集
从nginx服务器将日志数据传输到日志服务器
[root@VM_231_116_centos ~]# scp -r /mydata/code/deploy/nginx/logs 10.115.82.34:/mydata/logs
root@10.105.83.34's password:
access-power-20170929.log 100% 126KB 125.8KB/s 00:00
access-web-20171016.log 100% 2616KB 2.6MB/s 00:00
access-power-20170907.log 100% 1687KB 1.7MB/s 00:00
access-api-20170911.log 100% 1209KB 1.2MB/s 00:00
access-power-20170930.log 100% 1354KB 1.3MB/s 00:00
access.log 100% 45MB 45.2MB/s 00:00
access-api-20170907.log 100% 2960KB 2.9MB/s 00:00
access-power-20170906.log 100% 669KB 669.1KB/s 00:01
access-api-20170904.log 100% 9186KB 9.0MB/s 00:00
- 服务器之间文件（夹）复制
# 文件
scp local_file remote_username@remote_ip:remote_folder
或者
scp local_file remote_username@remote_ip:remote_file
# 目录
scp -r local_folder remote_username@remote_ip:remote_folder
4，日志解析
主要有几点：
- 逐行解析
- 正则匹配
- 日期的处理
- 批量写入数据库
# -*- coding: utf-8 -*-
import re
import time
import os
import arrow
import pandas as pd
import json
import io_tosql
import shutil
from sqlalchemy import create_engine
engine_user_info = create_engine(
"mysql+pymysql://{}:{}@{}:{}/{}".format('usr', 'pwd', 'host','port', 'db'),
connect_args={"charset": "utf8"})
def parse(filename):
month_abr = {"Jan":"01", "Feb":"02", "Mar":"03", "Apr":"04", "May":"05", "Jun":"06",
"Jul":"07", "Aug":"08", "Sep":"09", "Oct":"10", "Nov":"11", "Dec":"12"}
dfs = []
try:
i = 0
file = open(filename)
for line in file:
pattern = "(\d+\.\d+\.\d+\.\d+).*?\[(.*?)\].*?(\w+) (/.*?) .*?\" (\d+) \[(.*?)\] (\d+) \"(.*?)\" \"(.*?)\" \"(.*?)\""
s = re.search(pattern, line)
if s:
remote_addr = s.group(1)
local_time = s.group(2)
request_method = s.group(3)
request_url = s.group(4)
status = s.group(5)
request_body = s.group(6)
body_bytes_sent = s.group(7)
http_referer = s.group(8)
http_user_agent = s.group(9)
http_x_forwarded_for = s.group(10)
# 30/Sep/2017:01:08:39 +0000
for mon in month_abr.keys():
if mon in local_time:
local_time = local_time.replace(mon, month_abr[mon])
break
lt = arrow.get(local_time, "DD/MM/YYYY:HH:mm:ss")
lt = lt.shift(hours=8)
local_time = str(lt.datetime)
i = i+1
# print("line:{} > {}".format(i, local_time))
if request_body != '-':
try:
request_body = request_body.replace(r'\x22', '"').replace("null", '""')
request_body_dict = json.loads(request_body)
fund_id = request_body_dict.get('fund_id', None)
user_id = request_body_dict.get('user_id', None)
if user_id is None:
user_id = request_body_dict.get('userId', None)
except Exception as e:
print("request_body:{}".format(request_body))
print(e)
fund_id = None
user_id = None
else:
fund_id = None
user_id = None
if request_method not in ("GET", "POST"):
# print(request_method)
continue
df = pd.DataFrame({"remote_addr": [remote_addr], "request_method": [request_method], "local_time": [local_time],
"request_url": [request_url], "status": [status], "request_body": [request_body],
"body_bytes_sent": [body_bytes_sent], "http_referer": [http_referer],
"http_user_agent": [http_user_agent], "http_x_forwarded_for": [http_x_forwarded_for],
"fund_id": [fund_id], "user_id": [user_id]
})
df['create_at'] = time.strftime("%Y-%m-%d %H:%M:%S", time.localtime(time.time()))
# print(df)
dfs.append(df)
#每100条写数据库
if len(dfs) >= 100:
df_all = pd.concat(dfs)
df_all = df_all.drop_duplicates(subset=['remote_addr', 'request_url','local_time'])
df_all.to_sql("log_table", engine, if_exists="append", index=False)
print("写入长度为：" + str(len(df_all)))
dfs = []
df_all = pd.concat(dfs)
df_all = df_all.drop_duplicates(subset=['remote_addr', 'request_url','local_time'])
df_all.to_sql("log_table", engine, if_exists="append", index=False)
except Exception as e:
print(e)
5，日志展示
日志结构化写入数据库后，到前端页面可以多维度展示，下面是展示页面示例：
-
统计每日活跃IP数
-
统计每日API请求次数
-
分类分析
