Reorganize metric integration tests

[dashpole]

My motivation for this is to be able to link to individual fixtures for each metric type (gauge, sum, non-monotonic sum, etc) for the standard and GMP exporter in documentation.

This is probably easiest to review commit-by-commit.

• makes a function for the GMP config
• simplifies naming of fixture files (e.g. removes most instances of "metrics" and "basic" from names)
• adds GMP tests for simple scenarios.
• Enables the normalization feature gate for GMP tests. This demonstrates correct name translation in fixtures for the GMP exporter.
• Change units of metrics to UCUM so the translation is tested.
• Adds basic resource attributes to simple scenarios.
• Disables service metric labels for fixtures
• Change names of metrics and labels to include .s so that it tests sanitization.

[codecov]

Codecov Report

Merging #674 (d0d9275) into main (7d82f66) will increase coverage by 0.48%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main     #674      +/-   ##
==========================================
+ Coverage   69.05%   69.54%   +0.48%     
==========================================
  Files          36       36              
  Lines        4699     4699              
==========================================
+ Hits         3245     3268      +23     
+ Misses       1302     1278      -24     
- Partials      152      153       +1     

Impacted Files	Coverage Δ
...tor/integrationtest/testcases/testcases_metrics.go	100.00% <100.00%> (+14.47%)	⬆️

... and 2 files with indirect coverage changes

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[dashpole]

rebased

[damemi]

nothing blocking, but some suggestions.

it would also be nice if we added a comment on each test explaining what are the critical things to look for in the input/output fixtures. some are pretty self-explanatory but as they get more complex it gets harder to verify the expected fixtures

[dashpole]

clarified all of the test case names to make expectations clearer.