Add support for OTEL_EXPORTER_OTLP_METRICS_TEMPORALITY_PREFERENCE environment variable

[utpilla]

Fixes #3756

Spec: https://github.com/open-telemetry/opentelemetry-specification/blob/v1.14.0/specification/metrics/sdk_exporters/otlp.md#additional-configuration

Changes

• Add support for OTEL_EXPORTER_OTLP_METRICS_TEMPORALITY_PREFERENCE environment variable

Please provide a brief description of the changes here.

Merge requirement checklist

• CONTRIBUTING guidelines followed (nullable enabled, static analysis, etc.)
• Unit tests added/updated
• Appropriate CHANGELOG.md files updated for non-trivial changes

[codecov]

Codecov Report

Merging #4667 (9d3f89a) into main (67bd46d) will decrease coverage by 0.02%.
The diff coverage is 100.00%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #4667      +/-   ##
==========================================
- Coverage   85.08%   85.07%   -0.02%     
==========================================
  Files         314      314              
  Lines       12725    12747      +22     
==========================================
+ Hits        10827    10844      +17     
- Misses       1898     1903       +5     

Impacted Files	Coverage Δ
...nTelemetryProtocol/OtlpMetricExporterExtensions.cs	98.86% <100.00%> (+1.89%)	⬆️

... and 3 files with indirect coverage changes

[CodeBlanch]

nit: Use the constant OtlpMetricExporterExtensions.OtlpMetricExporterTemporalityPreferenceEnvVarKey here (I'm assuming the tests can see internals)

[utpilla]

This was done on purpose. If someone accidentally changes the environment variable key name in the src file, then our tests should complain about it. If I use the const string defined in the src file then our tests would simply pass even when the value of the key is changed.

[CodeBlanch]

I don't necessarily disagree with your logic, but it does break with all other envvar tests that I can recall seeing. Ex:

opentelemetry-dotnet/test/OpenTelemetry.Exporter.OpenTelemetryProtocol.Tests/OtlpExporterOptionsTests.cs

Lines 62 to 65 in e7f7cd6

	Environment.SetEnvironmentVariable(OtlpExporterOptions.EndpointEnvVarName, "http://test:8888");
	Environment.SetEnvironmentVariable(OtlpExporterOptions.HeadersEnvVarName, "A=2,B=3");
	Environment.SetEnvironmentVariable(OtlpExporterOptions.TimeoutEnvVarName, "2000");
	Environment.SetEnvironmentVariable(OtlpExporterOptions.ProtocolEnvVarName, "http/protobuf");

I would probably go with the constant to be consistent and follow DRY principle. If someone does have a valid reason to change it they should only have to do it in one spot.

But I made this a "nit" comment because I figured you might disagree 🤣 Up to you!

[utpilla]

I think we should think about what do we want to test? If we want to test, that a user a can configure something by providing a value for a particular string in the config, then our tests shouldn't be using the same const as the src file. Our tests would be too dynamic for their own good. 😄

[CodeBlanch]

LGTM with a CHANGELOG update

[vishweshbankwar]

Remove this todo as well. Can be a separate PR

[utpilla]

Good call out. I was thinking we should update that README once we have released a version of the package with this feature instead of doing that now.