Skip to main content

​
title: “Using continue-on-error in GitHub Actions” description: “Learn how to apply continue-on-error at step and job level to prevent non-critical failures from stopping your GitHub Actions workflows.”

GitHub Actions provides the continue-on-error keyword to help you manage non-critical failures in your CI/CD workflows. By default, a non-zero exit code stops the job and any downstream steps. With continue-on-error, you can:
  • Skip failing steps while continuing the job
  • Mark jobs as neutral instead of failed
The image shows a GitHub Docs page about GitHub Actions, specifically focusing on the "continue-on-error" feature in workflow syntax. It includes an example of preventing a specific failing matrix job from causing a workflow run to fail.

​
Example Scenario

Imagine a Code Coverage job that runs tests and then validates coverage. If coverage falls below your threshold, the job fails immediately: 
ERROR: Coverage for lines (88.88%) does not meet global threshold (90%)
File              % Stmts   % Branch   % Funcs  % Lines   Uncovered Line #s
All files         87.88      50         87.5    88.88     21,47-48,56
app.js            87.5       88.88     87.5     88.88     21,47-48,56
Error: Process completed with exit code 1.
Because of this exit code, the artifact upload step never runs.

​
Step-Level continue-on-error

To allow subsequent steps to execute even when coverage fails, add continue-on-error: true to the Check Code Coverage step: 
jobs:
  code-coverage:
    name: Code Coverage
    runs-on: ubuntu-latest
    steps:
      - name: Checkout Repository
        uses: actions/checkout@v4

      - name: Setup Node.js
        uses: actions/setup-node@v3
        with:
          node-version: 18

      - name: Install Dependencies
        run: npm install

      - name: Check Code Coverage
        continue-on-error: true
        run: npm run coverage

      - name: Archive Test Result
        uses: actions/upload-artifact@v3
        with:
          name: Code-Coverage-Result
          path: coverage
          retention-days: 5
After this change, the coverage step logs an error but the workflow proceeds: 
Run npm run coverage
► Solar [email protected] coverage
    nyc --reporter cobertura --reporter lcov --reporter text --reporter json-summary mocha app-test.js --timeout 10000 --exit

Server successfully running on port - 3000

Planets API Suite
  Fetching Planet Details
    ✓ it should fetch a planet named Mercury (34ms)
    … (other tests)

ERROR: Coverage for lines (88.8%) does not meet global threshold (90%)
Uploading artifact Code-Coverage-Result

​
Quick Comparison

ScopeApplies ToBehavior
Step-LevelIndividual stepOnly that step can fail silently
Job-LevelEntire jobJob marked neutral, continues all

​
Job-Level continue-on-error

For matrix-based workflows, you can skip failures across all steps by setting continue-on-error at the job level: 
jobs:
  code-coverage:
    runs-on: ${{ matrix.os }}
    continue-on-error: ${{ matrix.experimental }}
    strategy:
      fail-fast: false
      matrix:
        node: [13, 14]
        os: [macos-latest, ubuntu-latest]
        experimental: [false]
        include:
          - node: 15
            os: ubuntu-latest
            experimental: true
    steps:
      # same steps as above…
Setting fail-fast: false ensures that other matrix runs continue even if one instance fails.
When matrix.experimental is true, failures in that job are treated as neutral, allowing the workflow to complete.

​
Workflow Summary

In the summary below, unit tests pass, the code coverage job finishes with an error annotation, and the Code-Coverage-Result artifact is uploaded:
The image shows a GitHub Actions workflow summary with successful unit testing jobs and a code coverage job that completed with an error. It also lists artifacts produced during runtime, including "Code-Coverage-Result" and "Mocha-Test-Result."

Watch Video