Metricbeat: System Dashboard doesn't show data for the past hour (TZ issue?)

Metricbeat's `[Metricbeat System] Host overview ECS` dashboard doesn't show any data when I change the time interval to something like the last X minutes, altough I can confirm that the cluster is ingesting data from the beats (DevTools, Discover, Stack Monitoring).

It seens like a Timezone setting issue, because when I change the interval for some 1 or two hours in the past the dashboard work. I've already tryied to set Kibana's TZ to my TZ, but that didn't helped.

This is only happening to the `System` dashboards.

- Version: 7.6.2
- Operating System: Linux
- Discuss Forum URL: https://discuss.elastic.co/t/metricbeat-kibana-dashboard-with-almost-no-data/226760
- Steps to Reproduce:

1. Deploy an Elasticsearch 7.6.2 (or 7.6.1) cluster
2. Install Metricbeat 7.6.2 somewhere
3. ./metricbeat setup --dashboards
4. ./metricbeat -e

One of my `metricbeat.yml` files looks like this:

```
name: "laptop.acme.co"
tags: ["ACME", "facme.co"]
fields_under_root: true
fields:
cliente.id: 0
cliente.nome: "ACME"
cliente.dominio: "acme.co"

cloud:
id: "monitoramento:...=="
auth: "monitoramento:${MONITORAMENTO_PASSWORD}"

output.elasticsearch:
compression_level: 3

setup.template:
append_fields:
- name: cliente.id
type: short
- name: cliente.nome
type: keyword
- name: cliente.dominio
type: keyword

setup.ilm:
policy_name: "metricbeat-acme"
pattern: "{now/M{yyyy.MM}}-000001"
rollover_alias: "metricbeat-%{[agent.version]}-acme"

logging:
level: error
to_files: true
files:
path: /var/log/metricbeat
name: metricbeat
keepfiles: 7
permissions: 0644

monitoring:
enabled: true

metricbeat.modules:
- module: system
enabled: true
period: 1m
metricsets:
- cpu # CPU usage
- load # CPU load averages
- memory # Memory usage
- network # Network IO
- process # Per process metrics
- process_summary # Process summary
- uptime # System Uptime
- socket_summary # Socket summary
#- core # Per CPU core usage
#- diskio # Disk IO
- filesystem # File system usage for each mountpoint
#- fsstat # File system summary metrics
#- raid # Raid
- socket # Sockets and connection info (linux only)
- service # systemd service information

- module: docker
enabled: true
hosts: ["unix:///var/run/docker.sock"]
period: 1m
metricsets:
- container
- cpu
- diskio
- event
- healthcheck
- info
#- image
- memory
- network
```

Reproduce the report using the listed Elasticsearch 7.6.2 and Metricbeat setup steps, then inspect the [Metricbeat System] Host overview ECS dashboard with current and historical time ranges. Compare the dashboard's time handling with the data visible in Discover and DevTools; done means current data appears in the System dashboards without changing Kibana's timezone.

Written by the indexing model from the issue text.