Prevent Pekko callbacks from retaining request contexts

[amarziali]

What Does This Do

Fix a continuation leak in Pekko HTTP async handlers.

When an application Future completed with the request context active, the future returned to Pekko could run framework callbacks under that context. A Pekko actor envelope could then capture the request continuation and retain it indefinitely if the envelope was discarded.

The fix:

• returns a dedicated Promise to Pekko and completes it under the root context
• copies the Scala Try in completion-priority mode so attached context cannot reach Pekko
• ensures spans finish even when response decoration throws
• centralizes the Scala Promise completion-priority configuration
• adds deterministic regression tests for both propagation modes
• removes the corresponding flaky-test annotation

flowchart LR
  A["Application Future completes<br/>(request context active)"]

  A -->|"old transform"| B["Pekko callback captures<br/>request continuation"]
  B --> C["Discarded actor envelope<br/>retains continuation"]

  A --> D["Finish request span"]
  D --> E["Attach root context"]
  E --> F["Complete Promise exposed to Pekko"]
  F --> G["Pekko callback runs<br/>without request context"]

  classDef broken fill:#ffd6d6,stroke:#c62828
  classDef healthy fill:#d8f3dc,stroke:#2e7d32
  class B,C broken
  class D,E,F,G healthy

Loading

The implementation is based on Alexey K.'s existing unmerged alexeyk/pekko-fix work and has been adapted to the current codebase.

Motivation

Closes #9396

Additional Notes

Contributor Checklist

• Format the title according to the contribution guidelines
• Assign the type: and (comp: or inst:) labels in addition to any other useful labels
• Avoid using close, fix, or any linking keywords when referencing an issue
  Use solves instead, and assign the PR milestone to the issue
• Update the CODEOWNERS file on source file addition, migration, or deletion
• Update public documentation with any new configuration flags or behaviors
• Once approved, use merge queue to merge the PR

Jira ticket: [PROJ-IDENT]

[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 2026-09-29T07:49:34.584345Z	936eb21	PR opened
🔒 Security Review	✅ Completed 2026-09-29T07:49:24.176490Z	936eb21	PR opened

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.

[datadog-prod-us1-6]

More details

The Pekko-facing Promise is completed under the root context, while completion-priority mode copies the Scala Try before framework callbacks can inherit request context.

Was this helpful? React 👍 or 👎

_{Open Bits AI session}

_{🤖 Bits Code Review · Commit 936eb21 · @DataDog review to ask questions}

[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	14.03 s	13.96 s	[-0.1%; +1.2%] (no difference)
startup:insecure-bank:tracing:Agent	12.94 s	13.01 s	[-1.5%; +0.4%] (no difference)
startup:petclinic:appsec:Agent	17.13 s	16.78 s	[+1.2%; +3.0%] (significantly worse)
startup:petclinic:iast:Agent	17.02 s	17.06 s	[-1.0%; +0.6%] (no difference)
startup:petclinic:profiling:Agent	16.67 s	16.84 s	[-2.4%; +0.4%] (no difference)
startup:petclinic:sca:Agent	17.01 s	17.05 s	[-1.2%; +0.8%] (no difference)
startup:petclinic:tracing:Agent	16.15 s	16.17 s	[-1.1%; +0.9%] (no difference)

Commit: 1dd66fee · 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.

[sarahchen6]

Looks reasonable, thanks! Btw - the warning benchmark results are from the known AgentMeter benchmark (job). I re-ran to confirm, and it's good now (job).

[dougqh]

(Comment drafted by Claude on behalf of @dougqh.)

Question on scope: akka-http-10.0 has its own DatadogAsyncHandlerWrapper, and AkkaHttp2ServerInstrumentation still lists DatadogAsyncHandlerWrapper$1 and $2 as helpers, which suggests it still uses the same two-function transform pattern that this PR replaces for Pekko. If so, the Akka HTTP/2 async handler path would have the same continuation-retention leak.

Is leaving Akka out intentional, for example because it does not reproduce there or because you want to keep this PR small? If it is just a follow-up, it might be worth filing an issue so it does not get lost. I have not tried to reproduce the Akka leak, so this is a question rather than a confirmed bug.

[amarziali]

Is leaving Akka out intentional, for example because it does not reproduce there or because you want to keep this PR small? If it is just a follow-up, it might be worth filing an issue so it does not get lost. I have not tried to reproduce the Akka leak, so this is a question rather than a confirmed bug.

Yes I had a look to akka without finding the same evidence. So I left for follow up. Here I just would like to correct what I observed that is failing in CI (this is actually a flaky test behaviour that we'd like to fix and the diagnostic gave us few hint to understand and propose a fix)

[amarziali]

/merge

[gh-worker-devflow-routing-ef8351]

View all feedbacks in Devflow UI.

2026-10-01 12:10:55 UTC ℹ️ Start processing command /merge

---

2026-10-01 12:10:59 UTC ℹ️ MergeQueue: pull request added to the queue

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

---

2026-10-01 13:21:37 UTC ℹ️ MergeQueue: This merge request was merged