Course contentsShow
← Cloud and Infrastructure
Lesson 1710 of 2,98935. GitHub ActionsFree lesson

Artifacts: Upload and Download Between Jobs

Share build artifacts between jobs using upload-artifact and download-artifact actions.

Artifacts: Upload and Download Between Jobs

What you'll learn: How to share files and build outputs between jobs in GitHub Actions workflows using artifact upload and download actions.


Why Jobs Need to Share Files

In multi-job workflows, each job runs in a fresh, isolated environment. If Job A builds your application and creates a binary, and Job B needs to test that exact binary, you can't just reference it directly—the file is gone once Job A completes. Artifacts solve this by storing files temporarily so later jobs can retrieve them.

Think of artifacts like a shared locker at the gym: Job A finishes its workout (build), stores its belongings (build outputs) in the locker, and Job B can come by later with the key to grab them.

Uploading Artifacts

Use the actions/upload-artifact action to save files from a job:

jobs:
  build:
    runs-on: ubuntu-latest
    steps:
      - name: Build application
        run: make build
      
      - name: Upload build output
        uses: actions/upload-artifact@v3
        with:
          name: app-binary
          path: dist/app

The name identifies your artifact, and path specifies what to store (files or directories).

Downloading Artifacts

Later jobs use actions/download-artifact to retrieve stored files:

  test:
    needs: build
    runs-on: ubuntu-latest
    steps:
      - name: Download artifact
        uses: actions/download-artifact@v3
        with:
          name: app-binary
      
      - name: Run tests
        run: ./app --test

Artifacts are extracted into the current working directory by default. The needs: build ensures the test job waits for the build job to complete.


Key Takeaway: Use upload-artifact to save job outputs and download-artifact to retrieve them in dependent jobs, enabling seamless data sharing across your workflow's isolated execution environments.