Job dependencies & failure gates

Stop delivery when tests fail

Lesson 4 of 20 · 12 min · Module 2: Gates and build outputs

A failed prerequisite should block the change that depends on it.

Free. No installation: you write and run the code in your browser, and it is checked automatically.

By the end, you’ll be able to…

Model the safety dependency
Preserve failure as a useful signal

The idea

A workflow combines an event boundary, job dependencies, and ordered steps. Conditions and prerequisite outcomes determine which jobs can run.

jobs:
  check:
    runs-on: ubuntu-latest
    steps: [{run: npm test}]
  package:
    needs: check
    runs-on: ubuntu-latest
    steps: [{run: npm run build}]

01Model the safety dependency

A deployment should depend on the checks that justify it. Writing jobs in a particular order is not the same as declaring needs.

02Preserve failure as a useful signal

If tests fail, suppressing the error makes the pipeline tell a misleading story. The lab deliberately supplies a failing test outcome.

03Inspect what did not happen

The expected result is a failed test job and a skipped deploy job, with no new deployment recorded. A skipped unsafe action is successful gating.

Your turn

The lab makes npm test fail. Define test running npm test, and deploy needing test with run lab deploy notes:2.0. The deploy job must be skipped.

What the checks look for

Records the failing test
Blocks deployment through needs

Try it in the editor →

Take this with you

Treat the absence of an unsafe deployment as a behavior worth testing.

Go a little further

Change the lab test outcome and predict which jobs become eligible.