Timeouts in cluster monitoring

[juankaromo]

Wazuh	Elastic	Rev
3.9.4	7.x	527

Description
It seems that cluster-monitoring makes a disproportionate call for the data we are taking, since it returns a lot of alerts and fall. You have to look for graphs that consume less.
It doesn't make sense that 2 out of 10 times you give us a timeout only in this part of the App for that graph.

Steps to reproduce

1. Go to Management > Cluster monitoring
2. See timeout error (sometimes)

Screenshots

Additional context
We are using Kibana's own calls for monitoring /elastic/visualizations/cluster-monitoring.

[jesusgn90]

After some researching I've found a possible way to optimize this view and to reduce this load time:

Since the timelion visualizations don't need all the fields under _source, we are showing "counts", we can try to remove the _source from results.

This can be done setting the _source values in the searchSource instance as follows:

searchSource.setField('source',false)

Of course, we need to add some conditions to don't do this always, just in case we are feeding visualizations, but it's still needed when using the Discover itself.

[jesusgn90]

PR #1704 improves the load time, and we should not see more timeouts here for standard situations (all components are working)