Finding
testing-lab has no unit test CI workflow. The repository contains testable Python helper modules — tests/service_catalog/shared/kube.py and various pytest test files — but there is no pytest run in CI. The only CI workflow is Argo YAML linting (.github/workflows/lint.yaml).
This means:
- Python syntax errors in test helpers are not caught until runtime in a live cluster
- Logic bugs in
kube.py utilities (kubectl wrappers, pod selectors, HTTP probes) are never exercised offline
- Coverage cannot be tracked or gated
Recommendation
- Add
.github/workflows/unit-tests.yml that runs pytest tests/ -v on PRs and pushes to main
- Install test dependencies (
pytest, pytest-cov) in CI
- Add unit tests for
tests/service_catalog/shared/kube.py (mock subprocess.run to test kubectl wrappers offline)
- Optionally add a coverage threshold gate
Priority
- Impact: high — test helper regressions go undetected until live cluster runs
- Effort: low — a basic CI workflow and mock-based unit tests can be added quickly
Filed by quality agent (hold-gated mode)
Finding
testing-labhas no unit test CI workflow. The repository contains testable Python helper modules —tests/service_catalog/shared/kube.pyand various pytest test files — but there is nopytestrun in CI. The only CI workflow is Argo YAML linting (.github/workflows/lint.yaml).This means:
kube.pyutilities (kubectl wrappers, pod selectors, HTTP probes) are never exercised offlineRecommendation
.github/workflows/unit-tests.ymlthat runspytest tests/ -von PRs and pushes tomainpytest,pytest-cov) in CItests/service_catalog/shared/kube.py(mocksubprocess.runto test kubectl wrappers offline)Priority
Filed by quality agent (hold-gated mode)