使用Monit监控本地进程
目前用它监控某些服务,失败自动重启,同时监控特定的日志文件,如果有变化,就发邮件报警
安装不细写了,网上好多
我先用cat /proc/version看了下我的系统是el6的,于是wget http://pkgs.repoforge.org/monit/monit-5.5-1.el6.rf.x86_64.rpm
启动什么的就不管了,直接上配置文件:
set daemon 10 set logfile /var/monit/monit.log set idfile /var/monit/id set statefile /var/monit/state set mailserver smtp.xxxxxxxxx.com port 25 USERNAME "xxxxxxxx@xxxxxxxx.cn" PASSWORD "xxxxxxxx" set mail-format { from: xxxxxx@xxxxxxx.cn subject: monit alert -- $EVENT $SERVICE message: $EVENT Service $SERVICE Date: $DATE Action: $ACTION Host: $HOST Description: $DESCRIPTION } set alert xxxxxxxxx@xxxxxxxx.com with reminder on 3 cycles set httpd port 2812 and use address 192.168.1.45 allow xxxxx:***** allow @users readonly include /etc/monit.d/*
两个例子:
check process nginx with pidfile /var/run/nginx/nginx.pid start program = "/etc/init.d/nginx start" stop program = "/etc/init.d/nginx stop" if failed host localhost port 80 protocol http then restart
和文件变化的监控
check file test.log with path /tmp/test.log if changed size then alert
##########################################################
今天又测试了下监控目录变化:
从github上找了下代码,代码里的文档比官网上的细多了。。。。
check directory root with path /root if changed timestamp then alert alert ****@******.com
然后干脆把rocket的监控也加上了
check host 192.168.1.45 with address 192.168.1.45 #check system 192.168.1.45 # if failed icmp type echo count 1 with timeout 3 seconds then alert if failed host 192.168.1.45 port 9876 type tcp then alert if failed host 192.168.1.45 port 10911 type TCP then alert alert xxxxxxx@xxxxxxxxxx.com
##########################################################
#从http://blog.chinaunix.net/uid-21516619-id-1825022.html copy 过来的tomcat监控
check process tomcat with pidfile /var/run/catalina.pid # 这个要另外说明【2】
start program = "/etc/init.d/tomcat start" # 设置启动命令
stop program = "/etc/init.d/tomcat stop" # 设置停止命令
if 9 restarts within 10 cycles then timeout # 设置在10个监视周期内重
# 启了9次则超时,不再监视
# 这个服务。原因另外说明【3】
if cpu usage > 90% for 5 cycles then alert # 如果在5个周期内该服务
# 的cpu使用率都超过90%
# 则提示
# 若连续5个周期打开url都失败(120秒超时,超时也认为失败)
# 则重启服务
if failed url http://127.0.0.1:4000/ timeout 120 seconds for 5 cycles then restart
if failed url http://127.0.0.1:5000/ timeout 120 seconds for 5 cycles then restart
使用/var/run/catalina.pid这个pid文件来检查tomcat这个服务(服务名可以随便起),tomcat进程默认是不使用pid文件的,pid文件需要显式为tomcat设置,可以打开tomcat目录下的bin目录,打开catalina.sh文件,在开头(但不是第一行)处加入:
CATALINA_PID=/var/run/catalina.pid
即可指定pid文件,然后重启tomcat,这样就可以monit的配置中指定pid文件了。
参考:
http://blog.chinaunix.net/uid-24774106-id-3705291.html
http://blog.csdn.net/fan_hai_ping/article/details/41733783
http://blog.chinaunix.net/uid-21516619-id-1825022.html
https://github.com/arnaudsj/monit
http://mmonit.com/monit/