zabbix3.0.4 add monitoring of the process:
method One: Monitor with custom commands
Main idea: Through Ps-ef|grep sdk-push-1.0.0.jar |grep-v grep|wc-l This command to determine whether the process Sdk-push is running, if not equal to 1 means the process hangs, triggering the alarm
Monitored zabbix-agent side:
To add monitoring parameters:
# vim/etc/zabbix/zabbix_agentd.conf
Userparameter=sdk_push,ps-ef|grep Sdk-push-1.0.0.jar |grep-v grep|wc-l
Restart Zabbix-agent make configuration effective
Service Zabbix-agent Restart
Zabbix-server test is OK, when the port is not the default 10050, you need to add-p to specify the port:
Zabbix_get-s 1.1.1.1-p 20050-k Sdk_push
Add the appropriate monitoring to the Zabbix Web page:
Add Monitoring Item Item
Confuguration--Hosts--find the corresponding host, point to Items--Create item
Create a trigger on this monitoring item
Confuguration--Hosts--find the corresponding host, tap Triggers---Create Trigger
Of course, you can also define action actions to restart your app, which is not recommended if the application is more complex.
Finally do not forget the test, you can stop the process to see if the alarm
method Two: The specified process is monitored by the Zabbix's own template
The following is the monitoring configuration of the elasticsearch process, the parameter description in key, the first parameter is the process name, no need to fill in, it will make the monitoring is not accurate, the second parameter is the user name of the running process, the third is the status of the process, including: all (default), run, Sleep, Zomb, the fourth parameter is used to specify the characters contained in the process name and to filter the process.
Items--Create item
Name:amount of ES
Type:zabbix Agent
Key:proc.num[,yunva,all,elasticsearch]
Application:process Monitor
Key:proc.num[,yunva,all,elasticsearch] Value Yunva to be filled in according to the actual situation (user running Elasticsearch program)
Note Type to select the Type:zabbix agent because the Type:zabbix agent (Active) mode cannot execute remote commands
3. After configuring the monitoring item, add the trigger, the following trigger indicates that the last two times the value is 0, indicating that no elasticsearch process is running, then the alarm is set.
Trggiers--Create Trigger
Name:process of ES is down
Expression: {Hostname:proc.num[,yunva,all,elasticsearch].max (#2)}=0
Monitoring ports with your own templates add Item:name:yuyin file port 9999
Type:zabbix Agent
KEY:NET.TCP.LISTEN[9999]
Trigger
Name:yuyin Port 9999 is down
Expression: {Template OS Linux_yuyin_file_aiwaya:net.tcp.listen[9999].max (#2)}=0