进程监控-monit

一、monit安装

yum -y install pam* openssl flex openssl-devel
tar -xf monit.tar.gz && cd monit && ./configure && make && make install

二、用法参考

2.1、监控进程

#以进程PID方式监控
check process DD_CRON with pidfile /var/run/crond.pid
        start program = "/etc/init.d/crond start"
        stop  program = "/etc/init.d/crond stop"
#以进程名字方式监控
check process OPENSIPS matching "/data/opensips/sbin/opensips -P /var/run/opensips.pid -m 2048 -M 16"
        start program = "/data/opensips/sbin/opensipsctl start"
        stop program = "/data/opensips/sbin/opensipsctl stop"

2.2、监控端口

#监控端口，并以指定用户启动
#TCP端口
check host tomcat_service with address 127.0.0.1
        start program = "/etc/init.d/tomcat start" as uid "user" and gid "user" and with timeout 60 seconds
        stop  program = "/etc/init.d/tomcat stop" as uid "user" and gid "user" and with timeout 60 seconds
        if failed port 8088 with timeout 60 seconds
                then restart
        if 2 restarts within 2 cycles
                then exec "/usr/bin/python /home/shell-scripts/send_messages/send_meassage.py tomcat_service"
#UDP端口
check host devhost-udp with address XXX.XXX.XXX.XXX
    if failed port 3001 type udp then exec "/etc/init.d/app restart"

2.3、监控内存

#如果单进程内存使用超过 200M 则告警
check process NXLOG with pidfile /var/run/graylog/collector-sidecar/nxlog.pid
        start program = "/etc/init.d/collector-sidecar start"
        stop  program = "/etc/init.d/collector-sidecar stop"
        if totalmem > 200 MB then restart

2.4、监控磁盘

#监控磁盘使用率大于 65% 并告警
check filesystem ROOT with path /dev/vda1
        if space usage > 65% then alert

2.5、监控CPU和内存

#监控ping或icmp协议判断网络是否异常
check host example with address 0.0.0.0
       if failed icmp type echo count 3 with timeout 20 seconds then alert
       if failed ping then alert
       if failed port 5060 type udp protocol sip then alert

2.7、按时间段监控

#在星期1至星期天 凌晨0点至2点不做监控
check program PROC with path "/root/check_sip/check_status.sh"
        if status = 0 for 3 times within 5 cycles then exec "/bin/sh /root/call_mobile.sh"
        repeat every 4 cycles
        alert 648813099@qq.com with reminder on 4 cycles
        not every "* 0-2 * * 1-7"

2.8、自定义监控脚本

#检测web端口状态：
check program webstatus with path "/home/check_http_status.sh"
    if status = 0 then exec "/home/check_http_status.sh start"
    repeat every 5 cycles #(如果问题依然存在，在5次循环后，执行/home/check_http_status.sh start)

2.9、文件监控

check file http_log with path /var/log/httpd/error_log
    ignore match "doorduweb"
    ignore match "% Total"
    ignore match "Dload  Upload"
    ignore match "Permission denied"
    ignore match "favicon.ico"
    if match "[error]" then alert
    alert a@email.com
    alert b@email.com

四、参考文档

https://mmonit.com/monit/documentation/monit.html

https://mmonit.com/wiki/Monit/ConfigurationExamples#syslogd