What is the test pyramid and why does it matter?

Deep dive

Pyramid is about feedback loops:

• Unit tests run fast and isolate logic.
• Integration tests catch wiring issues.
• E2E tests cover critical user journeys only.

The goal is fast signal with minimal flakiness and maintenance.

Examples

Typical distribution:

70% unit / 20% integration / 10% E2E

Common pitfalls

• Inverted pyramid (too many E2E).
• Treating all logic as unit-only and missing integration bugs.
• Not aligning tests to risk (critical flows under-tested).

Interview follow-ups

• When would you intentionally add more E2E tests?
• How do you decide the right mix for a small team?
• How does the pyramid change for microservices?