Cherry-pick #15997 to 7.6: Correctly format the path to read data from npipe.

[ph]

Cherry-pick of PR #15997 to 7.6 branch. Original message:

What does this PR do?

This is not a fix for #15832 this will be tackle in another PR.

When http+npipe:// scheme was initially parsed by the url parser of
metric the generated PATH wasn't correctly setup up this mean that Metricbeat was not able to collect metrics over a npipe. So I've fixed the parsing and the tests.

Why is it important?

Fix an issue to read metric over a named pipe

Checklist

• My code follows the style guidelines of this project
  - [ ] I have commented my code, particularly in hard-to-understand areas
  - [ ] I have made corresponding changes to the documentation
  - [ ] I have made corresponding change to the default configuration files
• I have added tests that prove my fix is effective or that my feature works

Author's Checklist

• This need to be tested under Windows.

How to test this PR locally

Now to test the behavior on Windows you will need a beat that expose metrics over
named pipes, you can do so with a configuration like this.

metricbeat.modules:
  - module: system

output.console: ~

http.enabled: true
http.host: \\.\pipe\custom

After you can start a Metricbeat instance with the beat modules to
collect the metric information.

metricbeat.modules:
- module: beat
  hosts: http+npipe://./pipe/custom
  metricsets:
  - stats
  - state
output.console: ~

If you start metricbeat with the -e -d "*" flags this will output any
metrics fetched from the running beat instance.

Related issues

• Closes: Metricbeat cannot read beat expose metric via npipe #15938
• Relates: Npipe syntax to expose http monitoring is slightly different from consuming the metric #15832