Avoid to create new visualizations


(Roger) #1

Hi all,
I'm trying to use default Metricbeat visualizations and dashboard of System module.
Working just with one cluster I have not any issue but if I start to play with another cluster (i.e. another environment), I would like to avoid to import new visualizations (I know that I must have a different dashboard to manage different cluster). do you have any tips? is there a way to generate randomly the id's value of every visualization?
or is there a smart way to manage it?


(Jaime Soriano) #2

Hi @rschirin,

I think I don't completely understand your problem :thinking:

By having two clusters, do you mean that you have two different groups of hosts?

What is the problem with them? are they using different metricbeat versions? or what is the reason you want to setup the visualizations again?


(Roger) #3

for sure my question was not so clear. I will try to go deeper.
I have 2 servers, one for Test Environment and one for Production Environment. I want to collect their system metrics, so I decided to use Metricbeat with System module (just to start).
I configured metricbeat.yml for Test Env and started Metricbeat. I allowed also the import procedure into Kibana.
After that I configured also the other Metricbeat, more or less in the same way but different index.template value.
When I started the second Metricbeat (Prod Env), every visualizations and dashboard are overwritten.
so, is there a way to avoid it?
What I would like to get at the end is visualizations and dashboards for Test Env and visualizations and dashboards for Prod Env.


(Jaime Soriano) #4

What are the settings you are changing to modify the index template?
You could try to set setup.dashboards.index to some pattern that matches the indexes names of both environments (for example setting it to metricbeat-* if your indexes are metricbeat-prod-... and metricbeat-test-.... This way you'd still have an only set of dashboards, but they could be used for both sets of indexes.

At the moment there is no way to modify the dashboards ids, so you cannot have dashboards of different versions or with different patterns.

One thing you can try is Kibana Spaces, a new feature introduced in Kibana 6.5.0 that allows to have different environments for dashboards and visualizations, so you can have the same ids in different Spaces.
To use them with Beats, add the base URL to the setup.kibana.host setting. For example if you have an environment named test on a Kibana instance in http://kibana.example.com:5601 the config would be:

setup.kibana.host: 'http://kibana.example.com:5601/s/test'

Kibana Spaces requires x-pack basic, that is included in the default packages.


(Roger) #5

@jsoriano thank you .
I will try Spaces feature even if I'm facing a creation issue in this moment.
If I will not use it, have I to create import every time dashboard and visualizations with different _id and name, right?


(Jaime Soriano) #6

If you want to have multiple versions of the same dashboards, or the same dashboards with different indexes, then yes you'd need to use different ids.
But changing dashboard names or ids is not currently supported by beats, you'd need to modify the ids on your own. You may also want to open an enhancement request for your use case.