feat(openfeature): After Hook Exposure event writer

[eliottness]

What does this PR do?

This PR implements feature flag exposure tracking for the Datadog OpenFeature provider. It adds the ability to automatically capture and report
feature flag evaluations to Datadog's Event Platform Intake, providing visibility into which feature flags are being evaluated and by which users.

Key changes:

• Adds exposureWriter for buffering and flushing exposure events to the Datadog agent via EVP proxy
• Implements exposureHook that integrates with the OpenFeature SDK to capture evaluation details
• Introduces ProviderConfig field to allow configuration of exposure flush intervals
• Adds attribute flattening utilities to ensure subject attributes comply with EVP intake schema
• Updates documentation with configuration options and environment variables

Out of scope refactoring

Moved the tracer private method udsClient, defaultHTTPClient and defaultDialer to the internal package to simply code between telemetry/tracer/openfeature writers

Reviewer's Checklist

• Changed code has unit tests for its functionality at or near 100% coverage.
• System-Tests covering this feature have been added and enabled with the va.b.c-dev version tag.
• There is a benchmark for any new code, or changes to existing code.
• If this interacts with the agent in a new way, a system test has been added.
• New code is free of linting errors. You can check this by running ./scripts/lint.sh locally.
• Add an appropriate team label so this PR gets put in the right place for the release notes.
• Non-trivial go.mod changes, e.g. adding new modules, are reviewed by @DataDog/dd-trace-go-guild.

Unsure? Have a question? Request a review!

[pr-commenter]

Benchmarks

Benchmark execution time: 2025-11-04 13:34:07

Comparing candidate commit 7dff30e in PR branch eliottness/ffe-exposures with baseline commit b2e6ba7 in branch main.

Found 0 performance improvements and 0 performance regressions! Performance is the same for 24 metrics, 0 unstable metrics.

[eliottness]

@codex review

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[datadog-datadog-prod-us1]

⚠️ Tests

⚠️ Warnings

❄️ 11 New flaky tests detected

    Test from github.com/DataDog/dd-trace-go/v2/internal/orchestrion/_integration/99designs.gqlgen

TestLoadDefaultConfig from github.com/DataDog/dd-trace-go/v2/internal/orchestrion/_integration/aws.v2 (Datadog)

Failed

=== RUN   TestLoadDefaultConfig
    generated_test.go:19: Running setup
    lifecycle.go:79: 🐳 Creating container for image amazon/dynamodb-local:latest
    lifecycle.go:85: ✅ Container created: f9f947a088cc
    lifecycle.go:91: 🐳 Starting container: f9f947a088cc
    lifecycle.go:97: ✅ Container started: f9f947a088cc
    logs.go:29: Initializing DynamoDB Local with the following configuration:
        
...

    Test from github.com/DataDog/dd-trace-go/v2/internal/orchestrion/_integration/chi.v5

View all

🧪 14 Tests failed

    Test from github.com/DataDog/dd-trace-go/v2/internal/orchestrion/_integration/aws.v1

    Test from github.com/DataDog/dd-trace-go/v2/internal/orchestrion/_integration/confluent-kafka-go.v1

    Test from github.com/DataDog/dd-trace-go/v2/internal/orchestrion/_integration/confluent-kafka-go.v2

View all

_{This comment will be updated automatically if new data arrives.
🔗 Commit SHA: 7dff30e | Docs | Datadog PR Page | Was this helpful? Give us feedback!}

[eliottness]

@codex review

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[eliottness]

@codex review

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Avoid sending 256 empty exposure events on startup

The exposure writer initializes its buffer with make([]exposureEvent, 1<<8). That creates a slice with length 256, not just capacity, so before any flag is evaluated the periodic flush sees 256 entries and pushes them to the agent as zero‑valued exposures. On the first tick this will POST 256 bogus events (likely rejected by the agent and logged as errors) even though no exposures occurred. The buffer should start with length 0 and capacity 256 instead.

Useful? React with 👍 / 👎.