Hi forum.
Following the exploration of the ELK universe, now, since it is strongly encouraged around, I want to learn how to setup the 'Monitoring cluster' scenario. While documentation is quite detailed on every piece, I fail to see the overall picture.
So, I've created a second EL cluster mockup, with xpack, ssl/tls authentication, so on and so forth ... I got now two kibanas, on two mockup clusters. I'm still not started using metricbeat, so at the moment no metricbeat, no filebeat, just plain EL+Kibana.
I have enabled monitoring on both clusters through kibana (My uderstanding is that kibana did it for me easily, by using the Cluster update settings API to set the xpack.monitoring.collection.enabled to true) ... so I got the elasticsearch servers monitored (independently, each on its cluster) on each kibana, conected to each cluster.
Now my understanding is that, my goal should be to switch monitorng data from cluster A to cluster B ... questions rise to me:
- Would I, at the end, disable monitoring on cluster A? (asume cluster B is the 'Monitoring cluster') I mean ALL monitoring would be centrallized at B... isn't it?
- I guess I've to setup HTTP exporters in A pointing to B ... isn't it?
- What happens with local exporters? do the ones on A keep sending monitoring to itself? or once a HTTP exporters are setup Local ones theyŕe disabled? or should I disable them some how?
- What about monitoring data on cluster B ... is data from local exporters at B and incoming HTTP data from A treated equally, as monitorng data? or is monitoring data from A treated as 'production data' at B?
- I've readed that I may disable monitoring on B? should I? really? why? Isn't good to monitor both clusters health?
I know its lot of questions... but I find not a clear picture... hope you could help me clarify the scenario.
Thank you very much in advance.
Regards.