nagios 监控mysql（check_mysql_health插件）

nagios 监控mysql（check_mysql_health插件）

http://labs.consol.de/wp-content/uploads/2011/08/check_mysql_health-2.1.7.tar.gz

 

# mount /dev/cdrom /media/cdrom
# rpm -ivh /media/cdrom/Server/perl-DBI-1.52-2.el5.i386.rpm
# rpm -ivh /media/cdrom/Server/perl-DBD-MsSQL-3.0007-2.el5.i386.rpm --nodeps

 

# tar zxf check_mysql_health-2.1.7.tar.gz
# cd check_mysql_health-2.1.7
# ./configure
# make && make install

 

# mysql -u root -p123456
> grant all on *.* to ming@'localhost' identified by '123456';
> quit

 

注：让该用户有访问本地mysql的权限即可，可用mysql -u ming -p123456验证，权限不一定给all

 

方式一（localhost监控）：

 

# vi /usr/local/nagios/etc/objects/commands.cfg

 

define command{
        command_name    check_mysql_health
        command_line    $USER1$/check_mysql_health --hostname $ARG1$ --port $ARG2$ --username $ARG3$ --password $ARG4$ --mode $ARG5$
--warning $ARG6$ --critical $ARG7$
        }

:wq

# vi /usr/local/nagios/etc/objects/localhost.cfg

define service{
        use                             local-service         ; Name of service template to use
        host_name                       myhostname
        service_description             check_mysql_connections
        check_command                   check_mysql_health!192.168.1.10!3306!ming!123456!threads-connected!700!1000
        }

:wq

 

可先测试一下:

 

#./check_mysql_health --hostname 192.168.1.10 --port 3306 --username ming --password 123456 --mode threads-connected --warning 700 --critical 1000

 

注：在线程数（连接数）超过1000或无法连接时红灯报警，线程数在700以内0以上为绿灯，700到1000之间为黄灯。

 

方式二（nrpe监控）：

 

下面要借助于nrpe（nrpe安装步骤略）

# vi /usr/local/nagios/etc/nrpe.cfg

command[check_mysql]=/usr/local/nagios/libexec/check_mysql_health --hostname 192.168.1.10 --port 3306 --username ming --password 123456 --mode threads-connected --warning 700 --critical 1500

:wq

 

重启nrpe：

 

# killall nrpe
# /usr/local/nagios/bin/nrpe  -c /usr/local/nagios/etc/nrpe.cfg -d

# vi /usr/local/nagios/etc/objects/ming.cfg                     (ming.cfg为监控主机的文件，自建的)

define service{
        use                             local-service         ; Name of service template to use
        host_name                       myhostname
        service_description             check_mysql
        check_command                   check_nrpe!check_mysql
        }

:wq

 

 

附：

利用check_mysql_health实现多种监控模式：

       connection-time          (Time to connect to the server)
       uptime                   (Time the server is running)
       threads-connected        (Number of currently open connections)线程数
       threadcache-hitrate      (Hit rate of the thread-cache)慢查询
       slave-lag                (Seconds behind master)
       slave-io-running         (Slave io running: Yes)主从热备
       slave-sql-running        (Slave sql running: Yes)主从热备
       qcache-hitrate           (Query cache hitrate)
       qcache-lowmem-prunes     (Query cache entries pruned because of low memory)
       keycache-hitrate         (MyISAM key cache hitrate)
       bufferpool-hitrate       (InnoDB buffer pool hitrate)
       bufferpool-wait-free     (InnoDB buffer pool waits for clean page available)
       log-waits                (InnoDB log waits because of a too small log buffer)
       tablecache-hitrate       (Table cache hitrate)
       table-lock-contention    (Table lock contention)锁表率
       index-usage              (Usage of indices)
       tmp-disk-tables          (Percent of temp tables created on disk)
       slow-queries             (Slow queries)
       long-running-procs       (long running processes)
       cluster-ndbd-running     (ndnd nodes are up and running)
       sql                      (any sql command returning a single number)