Fix visibility of pipeline metrics along with reserved/duplicate node name detection

[anupdhml]

Pull request

Fix generation of pipeline metrics (as reported in #647) and improve handing of cases when a pipeline node is created with the same name as an existing one.

Description

Ensured consistent id for metrics node when it's added to the pipeline graph (as "metrics"). Added tests to verify pipeline and ramp metrics as well.

Also throw error and terminate pipeline creation, if a user happens to use id metrics for one of their trickle operators (earlier, this was allowed and we would stop getting metrics too) -- this is solved as part of general handling described in #650.

For later, we should rename the events measurement to reflect that it's pipeline specific (or merge metrics in ramp_events under events.)

Related

• Related Issues: fixes Monitoring : Pipeline metrics are not generated. #647, Terminate pipeline creation when a node already exists with the given name #650

Checklist

• The RFC, if required, has been submitted and approved
• Any user-facing impact of the changes is reflected in docs.tremor.rs
• The code is tested
• Use of unsafe code is reasoned about in a comment
• Update CHANGELOG.md appropriately, recording any changes, bug fixes or other observable changes in behavior

[coveralls]

Coverage decreased (-0.05%) to 80.046% when pulling c8f7e7c on fix-pipeline-metrics into 0f5af59 on main.

[mfelsche]

I think we should definitely throw an error if a user names a node in, out, err or metrics as these are reserved port, and thus pipeline operator names.

[anupdhml]

Opened #650 to generalize error reporting when a node already exists with the given name.

[mfelsche]

LGTM in general, I especially like the nice errors.
I just had a tiny question.