Keep runtime test mocks from breaking asynchronous logging

[abhi1693]

What changes and why?

Scheduler heartbeat and search-indexer shutdown fixtures replaced threading.Event on Python’s shared module. During backend tests, an asynchronous API log writer could then call isinstance(..., threading.Event) with a function instead of a class, lose its output thread and emit an unhandled-thread warning. This occurred in the #79 unit job.

Scope each fake to its service module’s threading binding. The heartbeat, wait, signal and cleanup assertions remain intact. A shared regression fixture now sends and drains a real queued log during every scheduler/indexer runtime case, and verifies that the writer remains alive.

Related issues

Related to #79. This unrelated fixture repair is isolated from that production refactor. Merge after #79 and before continuing #80–#84, one PR at a time.

Validation

• With the original global mocks, all six strengthened runtime cases failed and all six emitted thread exceptions when warnings were treated as errors.
• With scoped mocks, 89 scheduler, search-indexer and logging tests passed using -W error::pytest.PytestUnhandledThreadExceptionWarning.
• Ruff lint and formatting passed. Normal commit hooks passed, including full backend unit coverage, types and version checks. The 67.60% backend coverage minimum remains unchanged.
• Current-head CI passed 39 jobs, with 10 expected skips: CI required, all three CodeQL analyses, Sonar, dependency/secret/workflow checks, populated upgrades, real API/admin journeys, web/Chrome/Edge reader parity, performance comparison and ARM64 image checks.
• Hosted backend unit suite: 3,007 passed, no skips and no warnings, all 273 source files, 67.72% combined coverage. All 1,137 PostgreSQL/Redis integration tests, 818 web and 648 admin unit tests passed. The unhandled asynchronous logging warning present in Reduce backend duplication and strengthen service boundaries #79 is absent.
• Sonar MCP confirms gate OK, zero open issues and 0.0% new duplication. This test-only change has no new-production-code coverage condition. No quality rule or gate was changed.
• GitHub’s separate managed AI findings scan remains unavailable because the monthly quota is exhausted (HTTP 402); standard CodeQL passed. No paid/account changes were made.
• Changes are confined to backend test fixtures; there is no UI or runtime behavior change requiring screenshots or new product documentation.

Rollout notes

No application, dependency, database, workflow or deployment change. Auto-merge is disabled. Wait for current-head CI required, CodeQL and Sonar to pass and resolve all review threads before manually merging.

Ready for review

• The description explains the change and validation results, including gaps.
• Relevant regression tests are updated; product documentation is unchanged because application behavior is unchanged.

[gitar-bot]

Code Review ✅ Approved

🟢 Low risk · Test fixtures isolate thread mocks and exercise queued logging without changing runtime code.

Isolates runtime test mocks for scheduler heartbeat and search-indexer shutdown to prevent asynchronous logging from breaking. Scoped fakes to each service module's threading binding while preserving heartbeat, wait, signal and cleanup assertions, and adds a regression fixture that verifies the log writer thread remains alive. No issues found.

Review coverage

📋 Rules No rules evaluated

🧪 Functional validation Not enabled · Set up

Options

Auto-apply is off → Gitar will not commit updates to this branch.
Display: compact → Counting what did not apply, without listing it.

Comment with these commands to change the behavior for this request:

Auto-apply	Compact
gitar auto-apply:on	gitar display:verbose

_{Was this helpful? React with 👍 / 👎 | Gitar}

[github-actions]

PR reader API performance

Result: INCONCLUSIVE

6 inconclusive · 2 within threshold. Zero request failures across all six runs.

Base 626947d3643f0848b9c90024574a0a1d923cb412 → PR 691a0386da9994cc06e17f98bc6d1d7b040c954c

Three independent runs per revision, each on a fresh GitHub-hosted ARM64 runner. 16 users, 60 seconds/run, 1,000 synthetic articles, cache off, five PgBouncer slots.

Latency comparison

All latency values are in milliseconds; lower is better. Values are medians of three runs.

Route	Base p95	PR p95	Change	Result
All requests	77	79	+2 (+2.6%)	Inconclusive
Latest feed	100	110	+10 (+10.0%)	Inconclusive
Feed pagination	72	70	-2 (-2.8%)	Inconclusive
Article detail	54	59	+5 (+9.3%)	Within threshold
Source feed	87	71	-16 (-18.4%)	Inconclusive
Topic feed	85	74	-11 (-12.9%)	Inconclusive
Feed filters	67	61	-6 (-9.0%)	Inconclusive
Sources	32	34	+2 (+6.2%)	Within threshold
Topics	32	37	+5 (+15.6%)	Inconclusive

Additional latency and throughput metrics

Each cell shows base → PR. Latencies are in milliseconds.

Endpoint	p50	p99	Requests/s
Aggregated	25 → 25	120 → 120	30.65 → 30.91
/v1/feed [latest]	41 → 40	140 → 150	8.92 → 9.05
/v1/feed [page]	24 → 24	110 → 97	3.85 → 3.87
/v1/articles/[id]	15 → 16	110 → 95	6.00 → 6.32
/v1/feed [source]	25 → 24	130 → 96	2.23 → 1.87
/v1/feed [topic]	27 → 24	90 → 88	1.96 → 1.76
/v1/feed/options	20 → 19	86 → 88	2.62 → 2.65
/v1/sources	10 → 10	53 → 54	2.62 → 2.65
/v1/topics	11 → 11	78 → 68	2.62 → 2.65

Individual runs

Run	Requests	Failures	Exit code
Base 1	1720	0	0
Base 2	1712	0	0
Base 3	1738	0	0
PR 1	1717	0	0
PR 2	1754	0	0
PR 3	1732	0	0

How to interpret this report

Regressed / Improved: a p95 regression needs >20% and >20 ms increase in the medians and at least two head samples versus the base median. Improvements use the inverse threshold. Within threshold means the change did not meet both thresholds.

Inconclusive: a within-revision p95 range >35% of its median and >20 ms is too noisy for a confident latency verdict.

Inconclusive latency comparisons are non-blocking: the check passes without claiming an improvement. Confirmed regressions and test/report errors fail the PR gate. Throughput is descriptive: paced users do not measure maximum capacity. This uncached public API test does not establish production health, browser performance or worker capacity. Separate runners can differ in hardware or host load; repetitions and noise checks reduce but cannot eliminate that uncertainty.

Full HTML/CSV reports and logs · attempt 1

[github-actions]

Backend unit coverage

PASS · minimum combined coverage 67.60%.

Tests: 3007 passed, 0 failed, 0 skipped. Source files: 273.

Workspace	Lines	Branches	Combined
All backend code	71.57%	53.67%	67.72%
apps/admin-api	64.78%	35.76%	59.95%
apps/aggregator	73.60%	57.18%	69.86%
apps/api	68.94%	55.88%	66.48%
apps/article-enrichment-worker	80.00%	50.00%	71.43%
apps/cli	68.76%	42.93%	64.81%
apps/images-worker	80.00%	50.00%	71.43%
apps/mcp	79.90%	73.88%	78.77%
apps/notifications	83.33%	76.67%	81.75%
apps/search-indexer	85.48%	85.71%	85.53%
apps/source-discovery-worker	80.00%	50.00%	71.43%
apps/user-api	59.19%	32.90%	54.27%
packages/core	73.66%	56.49%	69.57%
packages/http	94.43%	86.47%	92.62%

Unit coverage includes every backend workspace. PostgreSQL/Redis integration tests run separately.

Commit: 691a038 · HTML, JSON, XML reports and job logs · attempt 1

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues
0 New dependency risks

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud