阅读中文版本

Instrument Unit Tests with a Coverage Threshold

Enable coverage instrumentation and set an enforceable threshold so new code cannot merge below the agreed bar.

export default {
  test: {
    coverage: {
      provider: 'v8',
      thresholds: { lines: 80, functions: 75, branches: 70 },
    },
  },
};

npx vitest run --coverage || echo 'coverage gate failed'

What this recipe does

Configure coverage collection with a threshold that fails CI below the target.

Coverage is measured but never enforced, so it drifts down release after release.

Prerequisites

• A test runner with coverage such as vitest or jest
• A CI pipeline
• An agreed coverage target

Procedure

1. Enable coverage with thresholds per metric.

export default {
  test: {
    coverage: {
      provider: 'v8',
      thresholds: { lines: 80, functions: 75, branches: 70 },
    },
  },
};

Expected result: A run below any threshold exits non-zero, failing CI.

2. Run coverage and confirm the gate.

npx vitest run --coverage || echo 'coverage gate failed'

Expected result: The summary reports each metric against its threshold and fails when below.

How to confirm it worked

Run the commands above and confirm every expected result in order. If a step does not produce the stated output, use the troubleshooting notes before continuing. A passing run shows a commit that removes a test drops coverage below the threshold and fails CI.

Troubleshooting

Legacy code drags the number down — Ratchet per-directory thresholds and raise them as areas improve.

Coverage inflated by trivial tests — Pair coverage with mutation testing for critical modules.

Limitations

• Coverage measures execution, not correctness.
• Thresholds can be gamed.
• Generated code skews metrics.