最近发生了几次web应用停止响应，最后虽然解决了问题，但应该给应用加上监控
?
检查apache, ps aux | grep httpd 发现apache进程已经有10个之多, 再通过curl直接检查8080端口的jboss,发现jboss已经停止响应,从而拖垮了apahe
?
shell掌握有限, 简单用curl进行应用的监控
?
监控机器列表文件：
server.list
server1
server2
server3
?
建立监控脚本： webstatus.sh
?
#!/bin/sh
monitor_dir=/home/admin/monitor/
if [ ! -d $monitor_dir ]; then
??? mkdir $monitor_dir
fi
cd $monitor_dir
web_stat_log=web.status
if [ ! -f $web_stat_log ]; then
?? touch $web_stat_log
fi
server_list_file=server.list
if [ ! -f $server_list_file ]; then
?? echo "`date '+%Y-%m-%d %H:%M:%S'` ERROR:$server_list_file NOT exists!" >>$web_stat_log
exit 1
fi
#total=`wc -l $server_list_file|awk '{print $1}'`
for website in `cat $server_list_file`
do
?? url="http://$website/app.htm"
?? server_status_code=`curl -o /dev/null -s -m 10 --connect-timeout 10 -w %{http_code} "$url"`
?? if [ "$server_status_code" = "200" ]; then
??????? echo "`date '+%Y-%m-%d %H:%M:%S'` visit $website status code 200 OK" >>$web_stat_log
?? else
??????? echo "`date '+%Y-%m-%d %H:%M:%S'` visit $website error!!! server can't connect at 10s or stop response at 10 s, send alerm sms ..." >>$web_stat_log
??????? echo "!app alarm @136xxxxxxxx? server:$website can't connect at 10s or stop response at 10s ..." | nc smsserver port &
?? fi
done
exit 0
?
主要是利用 curl -o /dev/null -s -m 10 --connect-timeout 10 -w %{http_code} "$url" 返回状态码是否200,如果10s没有返回200状态码，则发警报
?
最后让linux 定时执行脚本：
crontab -e
*/10 * * * * /home/admin/app/bin/webstatus.sh
这样每隔10分钟就会执行一次