[Website]: Improve documentation for Fleet -> remote Elasticsearch output

[laraMorenoIgle]

Type of issue

None

What documentation page is affected

https://www.elastic.co/docs/reference/fleet/remote-elasticsearch-output

What happened?

Several suggestions:

1. Write a note that states clearly that the Fleet Server managing the agents needs to have access to the Elasticsearch cluster configured as remote Elasticsearch output.

   Some more detail: during the configuration of the remote output, you need to get a service token on the remote cluster.
   It is important to note that the service token just give the Fleet server the rights to create the API key that will be used by the elastic agent for the ingestion into the remote elasticsearch cluster.
   So Fleet Server will connect to the remote Elasticsearch using the Service token to generate the API key.

2. Per previous point, the note at the bottom of the page https://www.elastic.co/docs/reference/fleet/remote-elasticsearch-output is not completely accurate.

 In some cases, the remote Elasticsearch output used for Elastic Agent data can be reached by the Elastic Agents but not by Fleet Server. In those cases, you can ignore the resulting unhealthy state of the output and the associated Unable to connect error on the UI.

The some cases makes the sentence vague and could lead readers to think that the connectivity is not required. In fact, the connectivity is still required, but in some scenarios with multiple fleet servers that error could be harmless as long as one of the fleet servers have access to the remote cluster to use the service token and setup everything.

Consider for example this text for the note:

The Fleet Server will attempt to test all the Elasticsearch Remote Outputs. 
If, for instance, you have 3 different Fleet Servers with different scopes, it can be expected that 1 Fleet server is able to reach out an Elasticsearch Remote Output but not the others. 
In that case, the Elasticsearch Remote Output will be marked as unhealthy. But this can be ignored as long as the Fleet server that needs to have access to such Elasticsearch Remote Output, does have it.

3. User may need to add ssl settings for Fleet Server to connect to remote Elasticsearch (that's missing in the doc completely).
   It is not clear in https://www.elastic.co/docs/reference/fleet/remote-elasticsearch-output how to do that.
   We could either add some screenshots there explaining how to do it, or, at least add some links pointing to the documentation:
   https://www.elastic.co/docs/reference/fleet/tls-overview#output-ssl-options
   Please, note that latest version do have the options, but older ones (for instance, 9.0.1 don't, and you need to add the ssl.certificate_authorities in Advanced YAML configuration)

4. Also please note that there was a bug fixed on 8.15.0 onwards, [Fleet] Remote Elasticsearch output do not support custom ssl certificate fleet-server#3490, that ignored the SSL settings in Advanced YAML configuration

Additional info

No response

[eedugon]

Summary and action plan: This doc is missing important details.

1. We need to clarify the requirement of connectivity between Fleet Server and the remote Elasticsearch output (it's mandatory), even if that output won't be used by the Fleet server for data transfer.

2. There's a note in the doc that feels incorrect because it leads users to think the connectivity is not required. We need to either remove the note or explain that in case of multiple Fleet Servers, at least 1 should be able to reach the remote Elasticsearch output

3. Remind users to add the SSL CA certificate associated with the remote Elasticsearch cluster in case the remote cluster is not using publicly-trusted certificates (this is a very common use case and explained in other docs, and it's important to highlight it in this doc too).

PS - note that the exact connectivity requirement might need to be clarified by devs or PMs -> We know the connectivity is required, but in case of multiple Fleet Servers it's not totally clear if the requirement is that at least 1 (any) fleet server needs connectivity, or if it has to be the fleet server (or servers) that handle policies and agents that make use of that remote Elasticsearch output