Update Logs + Infrastructure docs to specify shipper compatibility and index customization

[tbragin]

If you look at the forums, as well as support cases, shipper compatibility with the new Infrastructure and Logs UI is one of the most frequent questions around getting started.
https://discuss.elastic.co/c/infrastructure
https://discuss.elastic.co/c/logs

Looking at our documentation, we are missing sufficient detail:
https://www.elastic.co/guide/en/infrastructure/guide/current/install-infrastructure-monitoring.html

At a minimum, I think we should add information around:
1. The fact that shippers of versions 6.5 and above are require
2. That users can customize in kibana.yml which index(es) this Logs and Infrastructure UIs looks at.

cc @alvarolobato @makwarth

[ppf2]

See also: elastic/kibana#26059

[tbragin]

cc @DanRoscigno

[tbragin]

also cc @weltenwort since he's been answering a lot of these questions on Discuss and is working on the UI to improve the situation. I imagine the docs will need an update again in 6.6, when we introduce the Settings UI.

[DanRoscigno]

These are not sufficient:

1. The fact that shippers of versions 6.5 and above are require
2. That users can customize in kibana.yml which index(es) this Logs and Infrastructure UIs looks at.

We also need to specify:

3. Which metrics from which metricsets from which Metricbeat modules drive the colors of the Infrastructure UI for host, pod, and container
4. What settings have to be included in metricbeat.yml in addition to enabling the correct modules (Maybe just host: ${NODE_NAME} for kubernetes module and adding NODE_NAME to the env?)

@alvarolobato , I see you referenced this from another issue, if you are scheduling this please count me in, I would be glad to help as I will be running build candidates to create new screenshots and testing with live systems.

[simianhacker]

I'm looking at the page (on Master) linked above. It looks like #1 is mentioned at the top of the page, #2 is also covered. Do the changes just need to be back ported to the older versions (6.5+)?

@DanRoscigno For #3 would you like a section for each node type (host, container, pod), then a listing of the metrics (CPU, Memory, etc) and the fields and aggregations behind them? I only mention the aggregation because for fields like CPU Usage we take system.cpu.user.pct and system.cpu.system.pct and add them together.

@exekias or @ruflin I will need some input (or a PR to my eventual PR) for #4. I'm not as familiar with the requirements on that side of the system and would feel more comfortable if one of you provided that section.

[DanRoscigno]

Thanks @simianhacker , I had not noticed that there were metrics being combined. I think that would be great.

[simianhacker]

From what I can tell it looks like the only thing that needs to be addressed is #4? @DanRoscigno or @tbragin thoughts?

[DanRoscigno]

I thought about this some more, it is necessary information, but it is not safe to have the same info in multiple places because we won't keep it up to date. So, how about this:

If you are managing Kubernetes see Running on Kubernetes .
Take care to have the host: ${NODE_NAME} set for the kubernetes module and set the NODE_NAME environment variable as specified in the metricbeat-kubernetes.yaml reference file referenced on that page.

Is that OK @simianhacker ?

[simianhacker]

If you are managing Kubernetes see Running on Kubernetes .
Take care to have the host: ${NODE_NAME} set for the kubernetes module and set the NODE_NAME environment variable as specified in the metricbeat-kubernetes.yaml reference file referenced on that page.

@DanRoscigno Do you have a place in mind to put this?