AWS CloudWatch - properly parse multiple values of dimensionName & dimensionValue lists

[zroubalik]

A regression introduced in 2.15

Checklist

• I have verified that my change is according to the deprecations & breaking changes policy
• Tests have been added
• Changelog has been updated and is aligned with our [changelog requirements](https://github.com/kedacore/keda/blob/main/CONTRIBUTING.md#Changelog
• Commits are signed with Developer Certificate of Origin (DCO - learn more)

Fixes #6236

[zroubalik]

/run-e2e aws_cloudwatch
Update: You can check the progress here

[JorTurFer]

Does it make sense adding an e2e test to cover this feature? (or updating the current one to use multi-dimensions, I'll asume that if multi-dimension works single dimension will work too)

[zroubalik]

Does it make sense adding an e2e test to cover this feature? (or updating the current one to use multi-dimensions, I'll asume that if multi-dimension works single dimension will work too)

IMHO unit test is enough, but 🤷‍♂️

[JorTurFer]

The current test only verifies that there aren't parsing errors, but it doesn't check the parsing result. I mean, the parsing didn't fail before either. If you remove the parser change and execute the test, it'll pass too.
To ensure this with unit tests, we need to include a verification of the parsed metadata

[JorTurFer]

Basically, the (previous) value unsplitted is used

but there isn't parsing errors

That's why I said about adding support with e2e or testing the metadata result with unit tests

[zroubalik]

You are right, my previous version of test covered this, then I did some refactoring to include the new test into exisiting tests and missed this check, will add it.

🤦‍♂️

[zroubalik]

@JorTurFer ptal

[JorTurFer]

/run-e2e aws_cloudwatch
Update: You can check the progress here

[JorTurFer]

Thanks!