用shell+sendmail实现服务器监控报警小脚本

这种email报警脚本遍地都是，很多用的sendmail、postfix，感觉有些大材小用了；也有些用perl的net::smtp和authen::sasl模块发信的，不过我perl用的不好，老发出些莫名其妙的邮件来（比如if(a>1){print(a);}，最后邮件里的显示的是0.99……）；最后采用sendemail这个成型的perl程序发信报警，而实时监控部分回归shell，终于完成。

wget <a href="http://caspian.dotconf.net/menu/software/sendemail/sendemail-v1.56.tar.gz">http://caspian.dotconf.net/menu/software/sendemail/sendemail-v1.56.tar.gz</a>
tar zxvf sendemail-v1.56.tar.gz
cp sendemail-v1.56/sendemail /shell/check/
cat >> check.sh <<eof

文件--check.sh：

#!/bin/bash
checkmail() {
    /usr/bin/perl ./sendemail -f [e-mail] -t [e-mail] –cc [e-mail] [e-mail]
-u "$subject" -m "$data" -s smtp.mail.com -xu userid -xp password
    sleep 300
}
while true;do
    loadavg=`awk '{print $2}' /proc/loadavg`
    diskuse=`df |awk '/cache/{print $5}'`
    servrun=`netstat -pln|awk -f/ '/:80/{print $nf}'`

    ip=`ifconfig|awk '/cast/{print $2}'|awk -f: '{if(nr==1){a=$2}else if(nr==2){b=$2}}end{print b"-"a}'`
    data=`echo -e "ip:$ip\nloadavg/5min:$loadavg\tcacheuse%:$diskuse\tservice:$servrun"`
    diskper=`echo $diskuse|sed 's/%//'`

    num=`ps aux|grep check.sh|grep -v grep|wc -l`

    if [[ $num > 2 ]];then
        break 2
    fi

    if [[ $loadavg > 1.00 ]] && [[ $diskper > 90 ]];then
        subject="warning-$ip-loadavg-disk"
        checkmail
    else if [[ $loadavg > 1.00 ]];then
        subject="warning-$ip-loadavg"
        checkmail
    else if [[ $diskper > 90 ]];then
        subject="warning-$ip-disk"
        checkmail
    fi
    sleep 60
done
eof

完成，执行sh check.sh &> /dev/null即可。
报警邮件：
标题：warning-192.168.0.100-10.10.10.10-disk ip:192.168.0.100-10.10.10.10 loadavg/5min:0.38 cacheuse%:94% service:nginx