Follow-up: improve AI usage rate specs

Background

The following discussions from !151955 (merged) should be addressed:

• @ekigbo started a discussion:

  nitpick: @rcrespo3 WDYT moving the CODE_SUGGESTIONS_USAGE_RATE test case out of here since it's the special case? This will mean we can avoid having a conditional in the test block and drop the last column here, making it all a little bit less complex to reason about these tests.

• @ekigbo started a discussion:

  nitpick: @rcrespo3 I'm not convinced this is a great way to test this, atm we're only testing that the lead time metric doesnt render the trend indicator if it's value is 0. Additionally, we dont test that the indicators point in the correct direction for each metric.

  Seems unlikely, but in the case of a regression we might miss changes with the trend indicator, its a bit more verbose but i think we could use additional mock values for different states (no change, positive change, negative change)

  Additionally AFAICT we dont render a trend indicator for the CODE_SUGGESTIONS_USAGE_RATE metric 🤔

• @ekigbo started a discussion:

  nitpick: @rcrespo3 I think we can simplify these provider creators as well as the wrapper creation methods 🤔

• @ekigbo started a discussion:

  nitpick: @rcrespo3 similar with the change tests, I think it might be good to move these tests to a new block.