快网CDN日志统计脚本

公司使用的快网的CDN，但是快网本身并不提供日志的分析统计，我这里还没有用elk，无奈只能先使用shell对日志进行分析统计，作为运维组对业务情况的一个大致的了解。

脚本是对前一天的CDNLOG进行解压缩分析，最后将结果按类输出到指定目录，再将这些结果打包扔回CDNLOG目录，因为CDNLOG目录是要上传云存储的，一并保存。其实还有一个动作，是将这些内容作为参数传给python脚本给一个监控系统，出图便于查看，这里没有贴上。

同样作为留底，不喜勿喷。

#!/bin/bash

function line()
{
cols=`tput cols`

for t in `seq 1 ${cols}`
do
echo -n "="
done
}

function global_define()
{
yesterday=`date -d "-1day" +%F`
comp_path="/srv/cdnlog/downloads/${yesterday}/"
cdnlog_list=`ls ${comp_path} | sed 's#.gz##g'`
uncomp_path="/srv/cdnlog/uncomp/${yesterday}/"
user=`whoami`
backup_path="/srv/cdnlog/downloads/${yesterday}/"
}

function global_check()
{
if [ $user != root ];then
echo "Please use root ."
exit 1
fi
}

function cycle()
{
for n in $cdnlog_list
do
check
define
uncomp
segmentation
source_address
method
http_code
http_size
referer
retime
platform
tag
back
cdn_node
response_size
delete
done
}

function check()
{
if [ -d ${uncomp_path}${n}/ ];then
rm -rf ${uncomp_path}${n}/
fi
}

function define()
{
gz_file="${n}.gz"
log_file="${n}.log"
log_path="/srv/cdnlog/uncomp/${yesterday}/${n}/"
}

function uncomp()
{
mkdir -p ${log_path}
gunzip -c ${comp_path}${gz_file} > ${log_path}${log_file}
}

function segmentation()
{
log_file_count=`cat ${log_path}${log_file} | wc -l`

awk -vsour_addr=${log_path}sour_addr.tmp -vmethod=${log_path}method.tmp -vhttp_code=${log_path}http_code.tmp -vhttp_size=${log_path}http_size.tmp -vreferer=${log_path}referer.tmp  -vplatform=${log_path}platform.tmp -vcdn_node=${log_path}cdn_node.tmp -vresponse_size=${log_path}response_size.tmp '{print $1 > sour_addr ;print $6,$7 > method ;print $9 > http_code ;print $10 > http_size ;print $11 > referer ;print $(NF) > cdn_node ;print $(NF-1) > response_size}' ${log_path}${log_file}

cat ${log_path}${log_file} | egrep "FCACHE_HIT_DISK|FCACHE_HIT|FCACHE_MISS" | awk -F 'FCACHE_HIT_DISK|FCACHE_HIT|FCACHE_MISS' '{print $2}' > ${log_path}more.tmp

awk -vretime1=${log_path}retime1.tmp -vretime2=${log_path}retime2.tmp -vretime3=${log_path}retime3.tmp -vretime4=${log_path}retime4.tmp -vretime5=${log_path}retime5.tmp -vretime6=${log_path}retime6.tmp -vretime7=${log_path}retime7.tmp -vretime8=${log_path}retime8.tmp -vretime9=${log_path}retime9.tmp -vtag=${log_path}tag.tmp -vback=${log_path}back.tmp '{print $1 > retime1 ;print $2 > retime2 ;print $3 > retime3 ;print $4 > retime4 ;print $5 > retime5 ;print $6 > retime6 ;print $7 > retime7 ;print $8 > retime8 ;print $9 > retime9 ;print $10 > tag ;print $12 > back }' ${log_path}more.tmp

cat ${log_path}${log_file} | awk -F '[(|)]' '{print $2}' > ${log_path}platform.tmp
}

function source_address()
{
line >> ${log_path}sour_addr.log
echo "Source Address Top 30 :" >> ${log_path}sour_addr.log
echo "" >> ${log_path}sour_addr.log
cat ${log_path}sour_addr.tmp | sort -r | uniq -c | sort -nr | head -30 >> ${log_path}sour_addr.log
sour_addr_total=`cat ${log_path}sour_addr.tmp | sort -r | uniq -c | wc -l `
echo "" >> ${log_path}sour_addr.log
echo "Source Address Total : $sour_addr_total" >> ${log_path}sour_addr.log
line >> ${log_path}sour_addr.log
}

function method()
{
line >> ${log_path}method.log
echo "Method & URL Top 30 :" >> ${log_path}method.log
echo "" >> ${log_path}method.log
cat ${log_path}method.tmp | sed 's#"##g' | sort -r | uniq -c | sort -nr | head -30 >> ${log_path}method.log
line >> ${log_path}method.log
}

function http_code()
{
line >> ${log_path}http_code.log
echo "Http Code Top :" >> ${log_path}http_code.log
echo "" >> ${log_path}http_code.log
http_code_list=`cat ${log_path}http_code.tmp | sort -r | uniq -c | sort -nr | awk '{print $2}'`
cat ${log_path}http_code.tmp | sort -r | uniq -c | sort -nr >> ${log_path}http_code.log
echo "" >> ${log_path}http_code.log

for y in $http_code_list
do
count=`cat ${log_path}http_code.log | awk -vncode=$y '$2==ncode{print $1}'`
rate=`echo "$count $log_file_count" | awk '{printf ("%.2f\n",$1*100/$2)}'`
echo -e "HTTP $y Count: \t$count" >> ${log_path}http_code.log
echo -e "HTTP $y Rate: \t\t${rate}%" >> ${log_path}http_code.log
echo "" >> ${log_path}http_code.log
done

line >> ${log_path}http_code.log
}

function http_size()
{
size_max=`cat ${log_path}http_size.tmp | sort -nr | head -1`
size_min=`cat ${log_path}http_size.tmp | sort -nr | tail -1`
size_total_tmp=`cat ${log_path}http_size.tmp | awk '{sum+=$1}END{print sum}'`
size_total=`echo $size_total_tmp | awk '{printf ("%.0f\n",$1)}'`
size_avg=`echo "$size_total $log_file_count" | awk '{printf ("%.2f\n",$1/$2)}'`

line >> ${log_path}http_size.log
echo "" >> ${log_path}http_size.log
echo -e "Http Size Max : \t$size_max" >> ${log_path}http_size.log
echo -e "Http Size Min : \t$size_min" >> ${log_path}http_size.log
echo -e "Http Size Total : \t$size_total" >> ${log_path}http_size.log
echo -e "Http Size Avg : \t$size_avg" >> ${log_path}http_size.log
line >> ${log_path}http_size.log
}

function referer()
{
line >> ${log_path}referer.log
echo "Referer Top 30 :" >> ${log_path}referer.log
echo "" >> ${log_path}referer.log
cat ${log_path}referer.tmp | sed 's#"##g' | sort -r | uniq -c | sort -nr | head -30 >> ${log_path}referer.log
line >> ${log_path}referer.log
}

function retime()
{
sed -ie '/-/d' ${log_path}retime{1..9}.tmp

for i in `seq 1 9`
do
eval retime${i}=`cat  ${log_path}retime${i}.tmp | wc -l`
eval sum${i}=`cat ${log_path}retime${i}.tmp | awk '{sum+=$1}END{print sum}' | awk '{printf ("%.2f\n",$1)}'`
eval avg${i}=`eval echo "\\${retime${i}} \\${sum$i}" | awk '{printf ("%.2f\n",$2/$1)}'`
eval max${i}=`cat ${log_path}retime${i}.tmp | sort -nr | head -1`
eval min${i}=`cat ${log_path}retime${i}.tmp | sort -nr | tail -1`
done

for k in `seq 1 9`
do
line >> ${log_path}retime${k}.log
echo "Retime${k} :" >> ${log_path}retime${k}.log
echo "" >> ${log_path}retime${k}.log
echo -en "Max \t: \t" >> ${log_path}retime${k}.log
eval echo "\${max$k}" >> ${log_path}retime${k}.log
echo -en "Min \t: \t" >> ${log_path}retime${k}.log
eval echo "\${min$k}" >> ${log_path}retime${k}.log
echo -en "Avg \t: \t" >> ${log_path}retime${k}.log
eval echo "\${avg$k}" >> ${log_path}retime${k}.log
echo -en "Total : \t" >> ${log_path}retime${k}.log
eval echo "\${sum$k}" >> ${log_path}retime${k}.log
line >> ${log_path}retime${k}.log
done

sed -i 's#-nan#0.00#g' ${log_path}retime{1..9}.log
}

function platform()
{
iphone_count=`cat ${log_path}platform.tmp | egrep "iPhone" | wc -l`
ipad_count=`cat ${log_path}platform.tmp | egrep "iPad" | wc -l`
mac_count=`cat ${log_path}platform.tmp | egrep "Macintosh" | wc -l`
android_count=`cat ${log_path}platform.tmp | egrep "Android|Adr" | wc -l`
windows_count=`cat ${log_path}platform.tmp | egrep "Windows NT|Windows 98|Windows XP" | wc -l`
windows_mobile_count=`cat ${log_path}platform.tmp | egrep "Windows Phone" | wc -l`
java_count=`cat ${log_path}platform.tmp | egrep "java" | wc -l`
linux_count=`cat ${log_path}platform.tmp | egrep "X11; Linux|linux-gnu" | wc -l`

iphone_rate=`echo "$iphone_count $log_file_count" | awk '{printf ("%.2f\n",$1/$2*100)}'`
ipad_rate=`echo "$ipad_count $log_file_count" | awk '{printf ("%.2f\n",$1/$2*100)}'`
mac_rate=`echo "$mac_count $log_file_count" | awk '{printf ("%.2f\n",$1/$2*100)}'`
android_rate=`echo "$android_count $log_file_count" | awk '{printf ("%.2f\n",$1/$2*100)}'`
windows_rate=`echo "$windows_count $log_file_count" | awk '{printf ("%.2f\n",$1/$2*100)}'`
windows_mobile_rate=`echo "$windows_mobile_count $log_file_count" | awk '{printf ("%.2f\n",$1/$2*100)}'`
java_rate=`echo "$java_count $log_file_count" | awk '{printf ("%.2f\n",$1/$2*100)}'`
linux_rate=`echo "$linux_count $log_file_count" | awk '{printf ("%.2f\n",$1/$2*100)}'`
line >> ${log_path}platform.log
echo "Platform Info :" >> ${log_path}platform.log
echo "" >> ${log_path}platform.log
echo -e "iPhone count \t\t: $iphone_count" >> ${log_path}platform.log
echo -e "iPhone rate \t\t: ${iphone_rate}%" >> ${log_path}platform.log
echo "" >> ${log_path}platform.log
echo -e "iPad count \t\t: $ipad_count" >> ${log_path}platform.log
echo -e "iPad rate \t\t: ${ipad_rate}%" >> ${log_path}platform.log
echo "" >> ${log_path}platform.log
echo -e "Mac count \t\t: $mac_count" >> ${log_path}platform.log
echo -e "Mac rate \t\t: ${mac_rate}%" >> ${log_path}platform.log
echo "" >> ${log_path}platform.log
echo -e "Android count \t\t: $android_count" >> ${log_path}platform.log
echo -e "Android rate \t\t: ${android_rate}%" >> ${log_path}platform.log
echo "" >> ${log_path}platform.log
echo -e "Windows count \t\t: $windows_count" >> ${log_path}platform.log
echo -e "Windows rate \t\t: ${windows_rate}%" >> ${log_path}platform.log
echo "" >> ${log_path}platform.log
echo -e "Windows mobile count \t: $windows_mobile_count" >> ${log_path}platform.log
echo -e "Windows mobile rate \t: ${windows_mobile_rate}%" >> ${log_path}platform.log
echo "" >> ${log_path}platform.log
echo -e "Java count \t\t: $java_count" >> ${log_path}platform.log
echo -e "Java rate \t\t: ${java_rate}%" >> ${log_path}platform.log
echo "" >> ${log_path}platform.log
echo -e "Linux count \t\t: $linux_count" >> ${log_path}platform.log
echo -e "Linux rate \t\t: ${linux_rate}%" >> ${log_path}platform.log
line >> ${log_path}platform.log
}

function tag()
{
tag_total=`cat ${log_path}more.tmp | wc -l`
tag_suc=`cat ${log_path}more.tmp | awk '{print $10}' | grep "1" | wc -l`
tag_fail=`cat ${log_path}more.tmp | awk '{print $10}' | grep "0" | wc -l`
tag_suc_rate=`echo "$tag_suc $tag_total" | awk '{printf ("%.2f\n",$1/$2*100)}'`
tag_fal_rate=`echo "$tag_fail $tag_total" | awk '{printf ("%.2f\n",$1/$2*100)}'`

line >> ${log_path}tag.log
echo "Tag of Download Info :" >> ${log_path}tag.log
echo "" >> ${log_path}tag.log
echo -e "Download Success Count \t\t: $tag_suc" >> ${log_path}tag.log
echo -e "Download Success Rate \t\t: ${tag_suc_rate}%" >> ${log_path}tag.log
echo -e "Download Fail Count \t\t: $tag_fail" >> ${log_path}tag.log
echo -e "Download Fail Rate \t\t: ${tag_fal_rate}%" >> ${log_path}tag.log
line >> ${log_path}tag.log
}

function back()
{
back_total=`cat ${log_path}more.tmp | awk '{print $12}' | egrep -v "-"| wc -l`
back_rate=`echo "$back_total $log_file_count" | awk '{printf ("%.2f\n",$1/$2*100)}'`

line >> ${log_path}back.log
echo "Back Source Info :" >> ${log_path}back.log
echo "" >> ${log_path}back.log
echo -e "Back Source Count \t: $back_total" >> ${log_path}back.log
echo -e "Back Source Rate \t: ${back_rate}%" >> ${log_path}back.log
line >> ${log_path}back.log
}

function cdn_node()
{
line >> ${log_path}cdn_node.log
echo "CDN Node Top 30 :" >> ${log_path}cdn_node.log
echo "" >> ${log_path}cdn_node.log
cat ${log_path}cdn_node.tmp | sort -n | uniq -c | sort -nr | head -30 > ${log_path}cdn_node.log
line >> ${log_path}cdn_node.log
}

function response_size()
{
response_size_total=`cat ${log_path}response_size.tmp | awk '{sum+=$1}END{print sum}' | awk '{printf ("%.0f\n",$1)}'`
response_size_avg=`echo "$response_size_total $log_file_count" | awk '{printf ("%.2f\n",$1/$2*100)}'`

line  >> ${log_path}response.log
echo "Response Size Info :" >> ${log_path}response.log
echo "" >> ${log_path}response.log
echo -e "Response Size Total \t: $response_size_total" >> ${log_path}response.log
echo -e "Response Size Avg \t: $response_size_avg" >> ${log_path}response.log
line >> ${log_path}response.log
}

function delete()
{
rm -f ${log_path}*.tmp
rm -f ${log_path}*.tmpe
rm -f ${log_path}${log_file}
}

function backup()
{
tar zcf ${backup_path}statistics_${yesterday}.tar.gz ${log_path}*.log
}

function main()
{
global_define
global_check
cycle
backup
}

main

本文出自 “菜鸟东” 博客，请务必保留此出处http://radish.blog.51cto.com/5944322/1738296