test: deflake test-inspect-async-hook-setup-at-inspect

[christianaurichzm]

parallel/test-inspect-async-hook-setup-at-inspect failed in 4 PRs in the daily reliability report, on win2019, win2022 and ubuntu2404-x64, always with:

AssertionError [ERR_ASSERTION]: callFrames,reason,hitBreakpoints contains "asyncStackTrace" property

The test treats the response to Debugger.setAsyncCallStackDepth as proof that async stack recording is on. It is not: dispatched from a V8 interrupt, the agent defers enabling the async hook to an immediate that the subsequent pause
prevents from running. Sending the command after the first pause dispatches it from the nested message loop instead, where the hook is enabled inline.

Reproduction

The race window is normally a few microseconds wide. Widening it in the test script makes the failure reliably reproducible:

- const waiting = setInterval(() => { debugger; }, 50);
+ const waiting = setInterval(() => { const t = Date.now(); while (Date.now() - t < 8); debugger; }, 10);

	widened	as-is, 24 concurrent processes
before	59 / 60	18 / 9600 (0.19%)
after	0 / 60	0 / 9600

All measurements are Linux x64 on 16 cores, so the 24 processes are deliberately oversubscribed. All 18 baseline failures carried the signature above.

Unwidened, node-stress-single-test is weak evidence here: it runs iterations sequentially on an idle worker, where the rate is lower still, and even at the 0.19% measured under contention 100 iterations have an 83% chance of coming back
clean. Widened, it needs no concurrency at all - 38 / 40 failures running one process at a time on an idle machine, 34 / 40 pinned to a single core.

Scope

Deferring the hook is deliberate: calling into JS from a V8 interrupt is not safe (#65028). This corrects an assumption in the test, not the runtime. Whether the protocol should expose when the hook becomes active is a separate matter.

Refs: https://github.com/nodejs/reliability/blob/main/reports/2026-08-26.md
Refs: #65028

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.05%. Comparing base (9baabd4) to head (f7a8857).
⚠️ Report is 3 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #65584      +/-   ##
==========================================
- Coverage   90.07%   90.05%   -0.02%     
==========================================
  Files         751      751              
  Lines      254875   254875              
  Branches    48108    48118      +10     
==========================================
- Hits       229579   229532      -47     
- Misses      16466    16518      +52     
+ Partials     8830     8825       -5     

see 33 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[nodejs-github-bot]

CI: https://ci.nodejs.org/job/node-test-pull-request/76595/