I really like the graphing ability of zabbix, but I question its scalability limitations caused by the gui. I am considering setting up zabbix to monitor 500+ hosts running multiple services each. I have some requirements which I have been able to find solutions for in Zabbix. If anyone knows how to solve these issues with zabbix, please share!
Logging in to 500+ servers to stop a service will obviously not be a good solution. I have an automation tool to do this from a central location. I need to somehow tie zabbix in to this.
Automation:
If a zabbix monitored service shuts down because the service_shutdown script was run, I need to automatically disable alerts. Is there any way to do this in zabbix? Either by setting something on the edge clients, or sending something to the zabbix server?
If If I add a new service to 500 hosts, is there any way to automatically add this new service?
Using command line based monitoring tools, this is all pretty easy. Its just a matter of having scripts edit/clone/modify config files. The zabbix gui scares me.
I would really love to use zabbix. Please share any ideas you may have, or experience you may have with monitoring a large number of things in a scalable fashion.
Thanks!
Logging in to 500+ servers to stop a service will obviously not be a good solution. I have an automation tool to do this from a central location. I need to somehow tie zabbix in to this.
Automation:
If a zabbix monitored service shuts down because the service_shutdown script was run, I need to automatically disable alerts. Is there any way to do this in zabbix? Either by setting something on the edge clients, or sending something to the zabbix server?
If If I add a new service to 500 hosts, is there any way to automatically add this new service?
Using command line based monitoring tools, this is all pretty easy. Its just a matter of having scripts edit/clone/modify config files. The zabbix gui scares me.
I would really love to use zabbix. Please share any ideas you may have, or experience you may have with monitoring a large number of things in a scalable fashion.
Thanks!
Comment