[elastic_agent]: Inconsistent memory report on FleetUI and metric dashboard

### Integration Name

Elastic Agent [elastic_agent]

### Dataset Name

_No response_

### Integration Version

main, 2.6.2

### Agent Version

all

### Agent Output Type

elasticsearch

### Elasticsearch Version

all

### OS Version and Architecture

all

### Software/API Version

_No response_

### Error Message

The memory shown on FleetUI for an Elastic Agent is inconsistent with the memory show on the _[Elastic Agent] Agent metrics_ dashboard.

### Event Original

_No response_

### What did you do?

I looked the Elastic Agent memory consumption on FleetUI and on the _[Elastic Agent] Agent metrics_.

It seems Elastic Defend is missing to be taken into account on FleetUI. However there might be other factors affecting it as well.

### MacOS with Elastic Defend

### Linux with Elastic Defend

### What did you see?

An inconsistency on the memory usage for the Elastic Agent

### What did you expect to see?

A consistent memory usage for the Elastic Agent on both places

### Anything else?

The idea is to ensure a consistent report for the main use-cases and platforms:
[with elastic defend / without elastic defend] X [linux, mac, windows]