opentelemetry-sdk: fix explicit histogram aggregation to handle multiple explicit bucket advisories

[lseguy]

Description

#4434 merged a fix that allowed the ExplicitBucketHistogramAggregation created by the OTLP exporter to take into account histogram buckets advice.

However, unlike the DefaultAggregation, it does not allow setting independent bucket boundaries for multiple histograms. The first histogram will set the explicit boundaries based on the advice, but the explicit bucket boundaries of further histograms are ignored.

This keeps the behavior of the View's explicit settings taking precedence over the advisory, except when ExplicitBucketHistogramAggregation uses the default bucket boundaries.

Type of change

Please delete options that are not relevant.

• Bug fix (non-breaking change which fixes an issue)

How Has This Been Tested?

Please describe the tests that you ran to verify your changes. Provide instructions so we can reproduce. Please also list any relevant details for your test configuration

• tox -e py311-test-opentelemetry-sdk

Does This PR Require a Contrib Repo Change?

• Yes. - Link to PR:
• No.

Checklist:

• Followed the style guidelines of this project
• Changelogs have been updated
• Unit tests have been added
• Documentation has been updated

[linux-foundation-easycla]

The committers listed above are authorized under a signed CLA.

• ✅ login: lseguy / name: Louis Séguy (848865e, cd86beb, 6f42342, f0bb399, 29c3a03, 54fa322)
• ✅ login: emdneto / name: Emídio Neto (fe6491c)

[xrmx]

@lseguy any reason this is in draft?

[lseguy]

@lseguy any reason this is in draft?

I just had to sign the CLA and update the CHANGELOG.
I'd be happy to get your review if you don't mind, since you worked on #4361 and #4434.

[emdneto]

Right. Thank you for the fix.

That change fixes the bug while keeping the _DEFAULT_EXPLICIT_BUCKET_HISTOGRAM_AGGREGATION_BOUNDARIES in case boundaries are not set. I know we have this test, but can you write one as part of the integration test? Like: asserting if we don't set the explicit buckets advisory it will keep the default ones?

[lseguy]

That change fixes the bug while keeping the _DEFAULT_EXPLICIT_BUCKET_HISTOGRAM_AGGREGATION_BOUNDARIES in case boundaries are not set. I know we have this test, but can you write one as part of the integration test? Like: asserting if we don't set the explicit buckets advisory it will keep the default ones?

Added a new test case checking for default boundaries, let me know if that's not what you had in mind.

[emdneto]

That change fixes the bug while keeping the _DEFAULT_EXPLICIT_BUCKET_HISTOGRAM_AGGREGATION_BOUNDARIES in case boundaries are not set. I know we have this test, but can you write one as part of the integration test? Like: asserting if we don't set the explicit buckets advisory it will keep the default ones?

Added a new test case checking for default boundaries, let me know if that's not what you had in mind.

All good. Thank you.