Retry sidecar connect before probe install

[Leiyks]

Description

Workflow Automation • View in Workflow Automation

Motivation (WHY)

• test_extension_ci has recurring flakes in live debugger / dynamic instrumentation tests where probes are not installed in time.
• The failures align with a startup race window where the sidecar transport can still be unavailable when request initialization proceeds, which prevents timely RC polling / probe application.
• This has been a major source of master pipeline instability (for example pipeline 113108177).

Changes (WHAT)

• Added a sidecar readiness retry in ext/sidecar.c during ddtrace_sidecar_rinit():
  • If DDTRACE_G(sidecar) is not yet connected but ddtrace_endpoint is available, call ddtrace_sidecar_ensure_active() before submitting root span data.
• This keeps the existing startup path intact and only adds a targeted second-chance connection at request init, reducing probe-install race likelihood without broad timeout increases.

Testing (HOW verified)

• Environment limitations prevented full local CI reproduction:
  • podman compose networking is unavailable in this sandbox (missing CNI plugins), so request-replayer/test-agent services could not be started.
  • libdatadog submodule fetch is blocked (403 allowlist restriction), so full extension build/test execution is not possible here.
• Validation performed:
  • Confirmed affected probe-install/test helper paths and sidecar/RC initialization flow in source.
  • Verified the patch is minimal and scoped to sidecar reconnect timing in ddtrace_sidecar_rinit().
  • Confirmed final diff only changes ext/sidecar.c.

Relevant failing pipeline example:
https://gitlab.ddbuild.io/DataDog/apm-reliability/dd-trace-php/-/pipelines/113108177

Reviewer checklist

• Test coverage seems ok.
• Appropriate labels assigned.

---

PR by Bits - View session in Datadog

Comment @DataDog to request changes

[datadog-prod-us1-4]

View session in Datadog

Bits Dev status: ✅ Done

CI Auto-fix: Disabled | Enable

Comment @DataDog to request changes

[datadog-prod-us1-6]

I can only run on private repositories.

[datadog-prod-us1-4]

🎉 All green!

❄️ No new flaky tests detected
🧪 All tests passed

🎯 Code Coverage (details)
• Patch Coverage: 100.00%
• Overall Coverage: 60.72% (+0.04%)

_{This comment will be updated automatically if new data arrives.
🔗 Commit SHA: 4cd2cdc | Docs | Datadog PR Page | Give us feedback!}

[pr-commenter]

Benchmarks [ tracer ]

Benchmark execution time: 2026-05-15 14:33:44

Comparing candidate commit 4cd2cdc in PR branch dd/fix-sidecar-rinit-retry-di-probes with baseline commit 4352865 in branch master.

Found 1 performance improvements and 0 performance regressions! Performance is the same for 192 metrics, 1 unstable metrics.

scenario:SamplingRuleMatchingBench/benchGlobMatching3-opcache

• 🟩 execution_time [-631.615ns; -285.585ns] or [-4.891%; -2.211%]