Smoke Test

You might say

After a new version reaches testing, first check that home opens, login works, and the core API responds; if one fails, do not start full regression.

A smoke test uses a few critical checks after a new build to decide whether the system is worth deeper testing.A new version can open home, log in, and call a core API first; any failure blocks slower tests. It aims to decide “continue or stop” within minutes. Re-running every old feature and edge case is regression testing, not smoke testing.

Know first

Build Test Case

Smoke TestBuild Verification TestBVT

How a new version passes three key probes first

RELEASE CANDIDATEweb v2.14.0

PENDING

rc-214.staging.trip.local

TRIPSearch destinations

○GET / returns 200○Account signs in○GET /api/trips returns 200

Smoke checks only decide whether the build deserves deeper testing; they do not prove every detail.

What smoke testing asks first

Decide whether deeper work is worthwhile: If home cannot open or login wholly fails, dozens of detailed cases mostly create noise. Smoke tests select the few basic, critical capabilities that expose an obviously unusable version early.

A failure needs a blocking action: Record the version and symptom, then fix or replace the test-environment version. Passing smoke does not mean enough quality; it only means deeper testing is worthwhile.

Quick checkChoose the best answer

After a new build deploys, the home page will not open. What should the team do next?

Record the version and failure, block the full regression run, fix it, then restart from smoke testsRun every payment, filtering, and settings case first, then revisit the home pageDeclare that every historical feature has regressed

You can say this to an AI Agent

After every deployment to the test environment, run smoke tests: home opens, login works, and the core API responds. If any fails, stop later regression, record the version and evidence; continue only when all three pass.

Learn next

1Regression Test→2Staging→