~95% on SimpleQA (e.g. Qwen3.6-27B on a 3090). Supports all local and cloud LLMs (llama.cpp, Ollama, Google, ...). 10+ search engines - arXiv, PubMed, your private documents. Everything Local & Encrypted.
Python academiaanthropicarxivbravedeep-researchencryptionhome-automationhomeserverlocallocal-deep-researchlocal-llmmistralollamaopenaipubmedresearchresearch-toolretrieval-augmented-generationsearxngself-hosted
ci(docker-tests): split test_settings_routes_coverage.py into a serial step (#4397)
LearningCircuit0c9ff7bf2bae
diff --git a/.github/workflows/docker-tests.yml b/.github/workflows/docker-tests.ymlindex e96f328db1..6bc129004c 100644--- a/.github/workflows/docker-tests.yml+++ b/.github/workflows/docker-tests.yml@@ -173,7 +173,15 @@ jobs:cd tests/infrastructure_tests && npm ci- name: Run all pytest tests with coverage- # Tests are always blocking - failures should not be silently ignored+ # Tests are always blocking - failures should not be silently ignored.+ # test_settings_routes_coverage.py is excluded here and run serially+ # in a dedicated step below — its 94 authenticated_client tests+ # (each one: register + login + SQLCipher KDF + 10 Alembic migrations)+ # accumulate enough connection/FD pressure inside the assigned xdist+ # worker to deadlock it silently late in the run (the worker dies+ # without a timeout, its coverage data is lost, total coverage drops+ # under the 50% fail-under gate, the job fails). Same pattern as+ # fd_canary tests, same fix: serial, no '-n auto' for this file.env:OPENROUTER_API_KEY: ${{ secrets.OPENROUTER_API_KEY }}run: |@@ -190,8 +198,8 @@ jobs:${{ env.TEST_IMAGE }} \sh -c "python -m pytest tests/ \--cov=src \- --cov-report=xml:coverage/coverage.xml \- --cov-report=html:coverage/htmlcov \+ --cov-report= \+ --cov-fail-under=0 \--tb=short \--no-header \-q \@@ -201,7 +209,37 @@ jobs:--ignore=tests/ui_tests \--ignore=tests/infrastructure_tests \--ignore=tests/api_tests_with_login \- --ignore=tests/fuzz"+ --ignore=tests/fuzz \+ --ignore=tests/web/routes/test_settings_routes_coverage.py"++ # See the comment on the main pytest step above for why this file is+ # split out. Runs SERIAL (no '-n auto'), appends coverage data into+ # the main run's .coverage file, and emits the final xml + html+ # reports + the 50% fail-under check after both runs have contributed.+ - name: Run test_settings_routes_coverage.py (serial, with coverage)+ env:+ OPENROUTER_API_KEY: ${{ secrets.OPENROUTER_API_KEY }}+ run: |+ docker run --rm \+ -v "$PWD":/app \+ -e CI=true \+ -e GITHUB_ACTIONS=true \+ -e LDR_TESTING_WITH_MOCKS=true \+ -e LDR_DISABLE_RATE_LIMITING=true \+ -e PYTHONPATH=/app/src \+ -e OPENROUTER_API_KEY="${OPENROUTER_API_KEY}" \+ -w /app \+ ${{ env.TEST_IMAGE }} \+ sh -c "python -m pytest tests/web/routes/test_settings_routes_coverage.py \+ --cov=src \+ --cov-append \+ --cov-report=xml:coverage/coverage.xml \+ --cov-report=html:coverage/htmlcov \+ --cov-fail-under=50 \+ --tb=short \+ --no-header \+ -q \+ --durations=20"# fd_canary tests (#3816 FD-leak guards) are CPU-bound real-event-loop# spins. Under coverage tracing + '-n auto' contention they blow pastdiff --git a/changelog.d/+ci-split-settings-routes-coverage-serial.misc.md b/changelog.d/+ci-split-settings-routes-coverage-serial.misc.mdnew file mode 100644index 0000000000..fd7df33f08--- /dev/null+++ b/changelog.d/+ci-split-settings-routes-coverage-serial.misc.md@@ -0,0 +1 @@+CI: split ``tests/web/routes/test_settings_routes_coverage.py`` out of the parallel pytest run into a dedicated serial step (same pattern as the existing ``fd_canary`` step). The 94 tests in that file all use the ``authenticated_client`` fixture (register + login + SQLCipher KDF + 10 Alembic migrations per test), and under ``-n auto`` they accumulate enough connection / FD pressure inside the assigned xdist worker to deadlock it silently late in the run — the worker dies with ``[gw0] node down: Not properly terminated`` (no timeout printed), its coverage data is dropped, and the 50% fail-under gate then fails the job even though every test that completed actually passed. Skipping individual tests in this file just relocates the death (confirmed empirically on this PR's earlier iteration). The serial step keeps all 94 tests running, appends into the main run's ``.coverage`` data via ``--cov-append``, and runs the final coverage report + 50% gate after both contributions have merged. Defence-in-depth: pairs with #4393's 180s/thread timeout.