Guard Tomcat CoyoteAdapter advice against null-scope NPE (quick fix)

[dougqh]

What Does This Do

Guards the Tomcat CoyoteAdapter server advice against a null-scope NullPointerException that can occur if the paired enter-advice fails before assigning its @Advice.Local ContextScope.

TomcatServerInstrumentation.ContextTrackingAdvice.closeScope and TomcatServerInstrumentation.ServiceAdvice.closeScope both called .close() unconditionally on an @Advice.Local ContextScope. Both are guarded with if (scope != null) now.

Motivation

Error Tracking issue: https://app.datadoghq.com/error-tracking/issue/5ce30218-b43b-11f0-ac08-da7ad0900002

java.lang.NullPointerException
  at (redacted: 6 frames)
  at java.base/java.lang.VirtualThread.run(VirtualThread.java:460)

Logged via the agent's own suppressed-exception mechanism as:

Failed to handle exception in instrumentation for org.apache.catalina.connector.CoyoteAdapter

ContextTrackingAdvice.extractParent and ServiceAdvice.onService are both @Advice.OnMethodEnter(suppress = Throwable.class). If either throws before assigning its @Advice.Local ContextScope (e.g. DECORATE.extract(req), DECORATE.startSpan(...), or a request-attribute call throwing inside an app's custom Request/attribute wrapper), the exception is silently swallowed by ByteBuddy's suppression bytecode. CoyoteAdapter.service() then proceeds normally, and the paired @Advice.OnMethodExit closeScope unconditionally calls .close() on the still-null local — a second NPE that is also swallowed and logged, fully masking the original failure. Because both exceptions are handled by generated advice bytecode rather than real application frames, the visible stack trace is almost entirely redacted, which is why this surfaces as a bare VirtualThread.run frame with no other context.

A related but narrower instance of this same failure mode (parentContext == null specifically) was already fixed in #11968 (rootContext() fallback), but the underlying structural issue — closeScope not tolerating a null local — was never addressed, so any other exception in the enter-advice reproduces the same NPE.

Additional Notes

Added a direct JUnit 5 unit test (TomcatServerInstrumentationTest) calling the two closeScope advice methods with null and with a real ContextScope, since both are plain public static methods and ContextScope is an interface — no bytecode instrumentation harness or mocking framework needed.

Contributor Checklist

• Add an entry in docs/release_notes.md if this change affects the user-facing functionality. — N/A, internal fix.
• Verify code coverage of new lines
• Confirm this change does not add flakiness in tests

Jira ticket: [PROJ-IDENT]

[datadog-datadog-prod-us1]

🎯 Code Coverage (details)
• Patch Coverage: 100.00%
• Overall Coverage: 68.82% (+9.51%)

_{This comment will be updated automatically if new data arrives.
🔗 Commit SHA: 827831f | Docs | Give us feedback!}

[dd-octo-sts]

🟢 Java Benchmark SLOs — All performance SLOs passed

Suite	Status
Startup	🟢 pass

SLO thresholds are defined here based on automatically generated metrics. A warning is raised when results are within 5% of the threshold.

PR vs. master results

Scenario	Candidate	master	Δ (95% CI of mean)
startup:insecure-bank:iast:Agent	13.96 s	13.95 s	[-0.8%; +1.0%] (no difference)
startup:insecure-bank:tracing:Agent	12.84 s	12.97 s	[-1.8%; -0.2%] (maybe better)
startup:petclinic:appsec:Agent	17.68 s	17.68 s	[-0.8%; +0.8%] (no difference)
startup:petclinic:iast:Agent	17.46 s	17.62 s	[-1.6%; -0.2%] (maybe better)
startup:petclinic:profiling:Agent	17.33 s	17.42 s	[-1.7%; +0.7%] (no difference)
startup:petclinic:sca:Agent	17.82 s	17.63 s	[+0.3%; +1.9%] (maybe worse)
startup:petclinic:tracing:Agent	16.66 s	16.73 s	[-1.4%; +0.6%] (no difference)

Commit: 827831f5 · CI Pipeline · Benchmarking Platform UI

---

Load and DaCapo benchmarks can be triggered manually in the GitLab pipeline. Results will appear in the Benchmarking Platform UI after completion.

[dougqh]

@codex review

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-07T20:49:34.683214Z	827831f	Draft marked ready

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" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. Can't wait for the next one!

Reviewed commit: 827831f5ea

ℹ️ About Codex in GitHub

Codex has been enabled to automatically 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 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

[datadog-datadog-prod-us1]

More details

Both CoyoteAdapter exit-advice methods tolerate a null scope after suppressed enter-advice failures, while preserving cleanup for non-null scopes.

Was this helpful? React 👍 or 👎

_{Open Bits AI session}

_{🤖 Bits Code Review · Commit 827831f · @DataDog review to ask questions}

[amarziali]

thanks

[dd-octo-sts]

/merge

[gh-worker-devflow-routing-ef8351]

View all feedbacks in Devflow UI.

2026-10-08 12:58:06 UTC ℹ️ Start processing command /merge

---

2026-10-08 12:58:09 UTC ℹ️ MergeQueue: pull request added to the queue

The expected merge time in master is approximately 1h (p90).

---

2026-10-08 14:19:29 UTC ℹ️ MergeQueue: This merge request was merged