test: stabilize distributed sub-DAG cancellation test

[yottahmd]

Summary

• stabilize the distributed sub-DAG cancellation test by waiting for the parent agent to record the running sub-DAG instead of polling socket-based current status
• assert the child sub-DAG reaches running state before stopping the parent run
• verify both the parent run and child sub-DAG finish as aborted

Root Cause

The flaky Windows job timed out waiting for DAGRunMgr.GetCurrentStatus / IsRunning to observe the parent as running. The parent was active while the distributed child ran, but that socket-based check could miss the state long enough for the test to fail.

Testing

• go test ./internal/intg/distr -run 'TestCancellation_SubDAG/cancelPropagatesToSubDAGOnWorker' -count=10
• go test ./internal/intg/distr -run TestCancellation_SubDAG -count=1
• make bin
• go test -timeout=20m -json ./internal/intg/distr -count=1 > /tmp/dagu-intg-distr-after-bin.json

Summary by CodeRabbit

• Tests
  • Enhanced internal test coverage for distributed lifecycle management, including improved validation of cancellation synchronization and sub-DAG status verification during shutdown scenarios.

[coderabbitai]

Important

Review skipped

Auto incremental reviews are disabled on this repository.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: 5e5b979d-b61a-4f2c-a802-559e6ccc8dee

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

📝 Walkthrough

Walkthrough

This test file modification updates lifecycle cancellation testing to use an error-carrying channel and adds comprehensive sub-DAG run validation. The changes enhance test verification by checking in-process agent status before cancellation and asserting sub-DAG run status becomes aborted afterward.

Changes

Cohort / File(s)	Summary
Lifecycle Test Enhancement internal/intg/distr/lifecycle_test.go	Replaced done channel synchronization with error-carrying errCh, added sub-DAG run status verification before and after cancellation, including extraction of subRunID, creation of rootRef, and assertions on final sub-DAG abort state.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~22 minutes

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title 'test: stabilize distributed sub-DAG cancellation test' accurately describes the main change: improving the reliability of a test for distributed sub-DAG cancellation by replacing socket-based polling with deterministic agent status checks.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix-flaky-test-24508081260

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 2

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Inline comments:
In `@internal/intg/distr/lifecycle_test.go`:
- Around line 168-170: The test currently does a one-shot call to
f.dagWrapper.DAGRunMgr.FindSubDAGRunStatus and asserts the child is Aborted,
which is race-prone; change the assertion to poll until the terminal state is
observed (or a timeout) by repeatedly calling FindSubDAGRunStatus in a loop or
using a test helper like require.Eventually, handling temporary nil/err
responses and only asserting equality to core.Aborted after the poll succeeds;
target the call site that invokes FindSubDAGRunStatus (in lifecycle_test.go) and
replace the direct require.Equal with a retrying/wait-until check that fails the
test if the terminal state isn’t reached within a bounded timeout.
- Line 137: In the require.Eventually polling blocks, replace calls using
context.Background() with the test's cancellable agent.Context so the polling
can be cancelled; specifically call agent.Status(agent.Context) instead of
agent.Status(context.Background()) and agent.FindSubDAGRunStatus(agent.Context,
...) instead of FindSubDAGRunStatus(context.Background(), ...) inside the two
polling closures (the require.Eventually blocks) to ensure proper cancellation
of the polling paths.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: 8b575628-bba4-4fc1-8e23-70684412c57a

📥 Commits

Reviewing files that changed from the base of the PR and between 8914d0e and 7a16421.

📒 Files selected for processing (1)

• internal/intg/distr/lifecycle_test.go