當(dāng)服務(wù)器壓力比較大，跑起來(lái)很費(fèi)力時(shí)候。我們經(jīng)常做站點(diǎn)頁(yè)面優(yōu)化，會(huì)去查找那些頁(yè)面訪(fǎng)問(wèn)次數(shù)比較多，而且比較費(fèi)時(shí)。 找到那些訪(fǎng)問(wèn)次數(shù)高，并且比較耗時(shí)的地址，就行相關(guān)優(yōu)化，會(huì)取得立竿見(jiàn)影的效果的。 下面是我在做優(yōu)化時(shí)候，經(jīng)常用到的一段shell 腳本。 這個(gè)也可以算是，統(tǒng)計(jì)web頁(yè)面的slowpage 慢訪(fǎng)問(wèn)頁(yè)面，象mysql slowquery 。
log_format main '$remote_addr - $remote_user [$time_local] $request '
'"$status" $body_bytes_sent "$http_referer" '
'"$http_user_agent" "$http_x_forwarded_for" $request_time';
access_log /var/log/nginx/access.log main buffer=32k;
#!/bin/sh
export PATH=/usr/bin:/bin:/usr/local/bin:/usr/X11R6/bin;
export LANG=zh_CN.GB2312;
function usage()
{
echo "$0 filelog options";
exit 1;
}
function slowlog()
{
#set -x;
field=$2;
files=$1;
end=2;
msg="";
[[ $2 == '1' ]] field=1end=2msg="總訪(fǎng)問(wèn)次數(shù)統(tǒng)計(jì)";
[[ $2 == '2' ]] field=3end=4msg="平均訪(fǎng)問(wèn)時(shí)間統(tǒng)計(jì)";
echo -e "\r\n\r\n";
echo -n "$msg";
seq -s '#' 30 | sed -e 's/[0-9]*//g';
awk '{split($7,bbb,"?");arr[bbb[1]]=arr[bbb[1]]+$NF; arr2[bbb[1]]=arr2[bbb[1]]+1; } END{for ( i in arr ) { print i":"arr2[i]":"arr[i]":"arr[i]/arr2[i]}}' $1 | sort -t: +$field -$end -rn |grep "pages" |head -30 | sed 's/:/\t/g'
}
[[ $# 2 ]] usage;
slowlog $1 $2;
..................
