How to measure and improve test coverage in a FastAPI project

Code Coverage with pytest-cov

Code coverage measures which lines of your application are executed by tests. It is a quality signal — not a target. 80% coverage with meaningful tests beats 100% coverage with trivial ones.

Install and run

pip install pytest-cov

# Run with coverage
pytest --cov=app --cov-report=term-missing

# Generate HTML report
pytest --cov=app --cov-report=html
open htmlcov/index.html

Configure with .coveragerc

[run]
source = app
omit =
    app/migrations/*
    app/database.py
    */tests/*

[report]
fail_under = 75

fail_under fails the test run if coverage drops below the threshold — useful in CI pipelines to prevent coverage regressions.

What to prioritize

Focus coverage on: business logic in CRUD functions, authentication edge cases (expired token, wrong password, inactive user), validation error paths (422 responses), and custom middleware. Skip: auto-generated migration files, configuration modules, and trivially obvious routes.

# Branch coverage catches missed conditional paths
pytest --cov=app --cov-branch --cov-report=term-missing

Branch coverage (--cov-branch) reports whether both sides of conditionals were tested. A line can be covered while an if/else branch remains untested.