I am looking to feed some statistics from my logging systems into Zabbix. Among the stats is how many logs/min I received from each server, each program, etc.
I cannot know ahead of time all the possible values.
I prefer to use active mode (pushed from the agent)
I'm currently using 2.4 but am looking at moving to 3.0 if that makes any difference.
With DataDog this is trivial, define the metric logs_per_server and have the server name be a tag
As far as I know Zabbix doesn't allow this, but zabbix does support discovery
could I setup a discovery template and then each minute as I send the stats, first send a discovery message listing all the values I currently have? (note that values sent in the past may not appear in the current list)
Is there a better way to do this?
The end result that for each metric type, I want a stacked graph showing all the systems that sent logs to that log server. When a new server shows up on the network, I want the logs for it to show up on the graph (even if the system isn't monitored by Zabbix, or is monitored under a different name)
I cannot know ahead of time all the possible values.
I prefer to use active mode (pushed from the agent)
I'm currently using 2.4 but am looking at moving to 3.0 if that makes any difference.
With DataDog this is trivial, define the metric logs_per_server and have the server name be a tag
As far as I know Zabbix doesn't allow this, but zabbix does support discovery
could I setup a discovery template and then each minute as I send the stats, first send a discovery message listing all the values I currently have? (note that values sent in the past may not appear in the current list)
Is there a better way to do this?
The end result that for each metric type, I want a stacked graph showing all the systems that sent logs to that log server. When a new server shows up on the network, I want the logs for it to show up on the graph (even if the system isn't monitored by Zabbix, or is monitored under a different name)