feat(openfeature): emit server-side EVP flagevaluation

[leoromanovsky]

Motivation

PHP server-side flag evaluations need to emit aggregated EVP flagevaluation payloads through the native Rust/C and libdatadog sidecar path while preserving the existing OTel metric and exposure paths. The emitted events must match the live flageval-worker contract: schema-visible aggregation dimensions only, no OpenFeature reason, bounded context before buffering, and targeting_rule.key only when real targeting-rule metadata exists.

Changes

• Adds native PHP EVP flagevaluation aggregation using backend-visible dimensions: flag key, variant key, allocation key, runtime-default state, error message, real targeting-rule key when present, targeting key, and bounded context.
• Prunes evaluation context before it participates in aggregation keys or queued snapshots.
• Keeps OpenFeature reason out of EVP payloads and aggregation keys.
• Preserves visible variant/allocation/error dimensions when overflow folds into degraded events.
• Flushes the aggregated flagevaluation batch before exposure and OTel metric sidecar actions at request shutdown.
• Splits native-to-sidecar flagevaluation batches into bounded 512-event IPC chunks.
• Bumps libdatadog to 3350d4b55c0815143e656f334967e4e4b6dd11f5, which contains the shared flagevaluation sidecar delivery fixes.

Decisions

• EVP cardinality is defined only by fields the worker receives; reason is not a hidden aggregate key.
• Degraded events omit targeting key and context but keep the visible dimensions needed for backend counts.
• PHP keeps the sidecar delivery architecture instead of adding a direct EVP HTTP writer to the extension.
• System-test validation covers the production per-flag degradation threshold and exercises overflow past the full-fidelity cap.
• This PR remains draft until the cross-SDK rollout and review pass are complete.

Validation Evidence

Validated with PHP SDK commit 9093fb9439fb34a69b0926df3c315435557e9abc, libdatadog commit 3350d4b55c0815143e656f334967e4e4b6dd11f5, and the merged system-tests flagevaluation baseline plus the local singular-path/overflow contract fix.

• RUSTC_BOOTSTRAP=1 cargo check -p datadog-php - PASS.
• RUSTC_BOOTSTRAP=1 cargo test -p datadog-php ffe -- --nocapture - PASS, 23 passed.
• ./tooling/bin/build-debug-artifact gnu-aarch64-8.0-nts /Users/leo.romanovsky/gsd-workspaces/flag-evaluations-cross-sdk/system-tests-clean-php/binaries - PASS; produced dd-library-php-1.21.0-aarch64-linux-gnu.tar.gz.
• SYSTEM_TEST_BUILD_TIMEOUT=1800 ./build.sh --library php --weblog-variant apache-mod-8.0 from system-tests-clean-php - PASS on 2026-06-20.
• TEST_LIBRARY=php WEBLOG_VARIANT=apache-mod-8.0 ./run.sh +l php FEATURE_FLAGGING_AND_EXPERIMENTATION -k "test_flag_eval_evp" - PASS on 2026-06-20, 8 passed, 2627 deselected in 60.08s with Library: php@1.21.0.
• Degradation payload evidence from the passing system-test run: payloads=30, events=10195, total evaluation_count=12708; evp-degradation-flag emitted 10001 rows with total evaluation_count=12000, including one degraded row with evaluation_count=2000 and omitted targeting_key/context.
• Real-backend dogfooding used local PHP builds for app-php7 and app-php8-openfeature with DD_EVP_PROXY_CONFIG_ADDITIONAL_ENDPOINTS={}; both returned 5/5 successful evaluations for ffe-dogfooding-string-flag.
• Real-backend targeting-key prefixes for EVP lookup:
  • server-evp-e2e-php7-1781931066-user-
  • server-evp-e2e-php8-openfeature-1781931066-user-
• Mock intake stayed empty during real-backend dogfooding: { "total_count": null, "by_path": {} }.

[datadog-official]

 

✨ Fix all issues with BitsAI

⚠️ Warnings

🚦 17 Pipeline jobs failed

DataDog/apm-reliability/dd-trace-php | ASAN test_c with multiple observers: [8.3]     

DataDog/apm-reliability/dd-trace-php | check-big-regressions     

DataDog/apm-reliability/dd-trace-php | test_extension_ci: [8.1]     

View all 17 failed jobs.

ℹ️ Info

No other issues found (see more)

🧪 All tests passed
❄️ No new flaky tests detected

🎯 Code Coverage (details)
• Patch Coverage: 100.00%
• Overall Coverage: 54.08% (-0.04%)

Useful? React with 👍 / 👎

_{This comment will be updated automatically if new data arrives.
🔗 Commit SHA: 9093fb9 | Docs | Datadog PR Page | Give us feedback!}

[pr-commenter]

Benchmarks [ tracer ]

Benchmark execution time: 2026-06-20 02:12:53

Comparing candidate commit 9093fb9 in PR branch leo.romanovsky/ffl-2446-evp-flagevaluation-php with baseline commit 8f132ce in branch master.

Some scenarios are present only in baseline or only in candidate runs. If you didn't create or remove some scenarios in your branch, this maybe a sign of crashed benchmarks 💥💥💥
Check Gitlab CI job log to find if any benchmark has crashed.

Scenarios present only in candidate:

• FlagEvaluationBench/benchEvaluateSplit
• FlagEvaluationBench/benchEvaluateDistinctContexts
• FlagEvaluationBench/benchEvaluateTargetingMatch-opcache
• FlagEvaluationBench/benchEvaluateWithoutCounting
• FlagEvaluationBench/benchEvaluateTargetingMatch
• FlagEvaluationBench/benchEvaluateDistinctContexts-opcache
• FlagEvaluationBench/benchEvaluateWithoutCounting-opcache
• FlagEvaluationBench/benchEvaluateSplit-opcache

Found 0 performance improvements and 1 performance regressions! Performance is the same for 193 metrics, 0 unstable metrics.

Explanation

This is an A/B test comparing a candidate commit's performance against that of a baseline commit. Performance changes are noted in the tables below as:

• 🟩 = significantly better candidate vs. baseline
• 🟥 = significantly worse candidate vs. baseline

We compute a confidence interval (CI) over the relative difference of means between metrics from the candidate and baseline commits, considering the baseline as the reference.

If the CI is entirely outside the configured SIGNIFICANT_IMPACT_THRESHOLD (or the deprecated UNCONFIDENCE_THRESHOLD), the change is considered significant.

Feel free to reach out to #apm-benchmarking-platform on Slack if you have any questions.

More details about the CI and significant changes

You can imagine this CI as a range of values that is likely to contain the true difference of means between the candidate and baseline commits.

CIs of the difference of means are often centered around 0%, because often changes are not that big:

---------------------------------(------|---^--------)-------------------------------->
                              -0.6%    0%  0.3%     +1.2%
                                 |          |        |
         lower bound of the CI --'          |        |
sample mean (center of the CI) -------------'        |
         upper bound of the CI ----------------------'

As described above, a change is considered significant if the CI is entirely outside the configured SIGNIFICANT_IMPACT_THRESHOLD (or the deprecated UNCONFIDENCE_THRESHOLD).

For instance, for an execution time metric, this confidence interval indicates a significantly worse performance:

----------------------------------------|---------|---(---------^---------)---------->
                                       0%        1%  1.3%      2.2%      3.1%
                                                  |   |         |         |
       significant impact threshold --------------'   |         |         |
                      lower bound of CI --------------'         |         |
       sample mean (center of the CI) --------------------------'         |
                      upper bound of CI ----------------------------------'

scenario:SpanBench/benchOpenTelemetryInteroperability

• 🟥 execution_time [+195.869µs; +199.038µs] or [+100.859%; +102.491%]