[Request]: Rework APM decision tree diagram

[colleenmcginnis]

Description

Rework the APM decision tree diagram based on feedback from @eedugon:

Great doc!! I have a question about this box:

Do you want to use other integrations sounds too generic when it probably wants to ask Do you want to use other integrations together with the APM Server or in the same APM Server?

I have a couple of extra questions, just in case you consider them useful, feel free to ignore them if wanted ;)

What does it mean centrally and on the edge in the context of the question? I haven't seen any explanation about that in the rest of the doc. It feels like on the edge means where the apps are running, far away from Elastic Stack, and centrally might mean then closer to the elastic stack, but I feel a bit lost with both answers.

See initial feedback from @simitt in #4599 (comment).

Resources

#4599 (comment)

Which documentation set does this change impact?

Stateful only

Feature differences

N/A

What release is this request related to?

N/A

Collaboration model

The documentation team

Point of contact.

Main contact: @eedugon

Stakeholders: @simitt

[eedugon]

About central and edge concepts, @simitt explained to me that:

• on the edge means multiple APM servers in the edge servers where applications run (so the applications have direct and faster access to APM server). In this scenario there's usually no connectivity between users browsers and APM servers, that's why this architecture wouldn't be compatible with RUM.

• central means running APM server in a location accessible by apps and users browsers, which is compatible with RUM.

In my opinion, the entire question about central and edge is not related with the decision of running the agent with the binary or as fleet managed, as in both options (on the edge and centrally) the user could run APM server binaries or Fleet managed agents).

It's true that on the edge means multiple APM servers and it's probably better from operational point of view to run them as managed (so the configuration could be changed in all at once directly from Fleet).

But none of that is obvious to the user or is explained in the docs.

Maybe we could add a (*) in the on the edge arrow explaining what on the edge means and why we recommmend Fleet Managed on that scenario, not being mandatory.