Link Search Menu Expand Document
LZone

Github Cheat Sheet

Actions

Environment variables

  • GITHUB_REF
  • GITHUB_BASE_REF
  • GITHUB_HEAD_REF
  • GITHUB_ENV (writable environment file passed between jobs)
  • GITHUB_WORKSPACE
  • GITHUB_REPOSITORY
  • GITHUB_REPOSITORY_OWNER (user or organisation name)
  • GITHUB_RUN_NUMBER (pipeline unique)
  • GITHUB_RUN_ID (globally unique)

More: https://docs.github.com/en/actions/learn-github-actions/environment-variables#default-environment-variables

Using workflow commands (annotations)

Documentation: https://docs.github.com/en/actions/using-workflows/workflow-commands-for-github-actions

echo '::set-output name={name}::{value}'

echo "::add-mask::MySecretPassword"

echo '::notice::My message'
echo '::error::My error'
echo '::error file={name},line={line},endLine={endLine},title={title}::{message}'

echo '::group::My group title'
echo '::endgroup::'

echo '::save-state name={name}::{value}'     # Export env vars to pre/post workflow actions

Interaction/input fields on manually triggering workflow

name: someworkflow
on:
  workflow_dispatch:
    inputs:
      branch:
        description: Please enter branch name
        default: main
      text:
        description: Describe your mood today
        required: true
        
steps:
  - name: Use input
    run: BRANCH="${{ github.event.inputs.branch }}" ./run "${{ github.event.inputs.text }}"

Switching runners using output variables

jobs:
  init:
    # Running an init step on a default runner to autosense the code
    outputs:
      mode: ${{ steps.autosense.outputs.runner }}    # Mandatory declaration
    steps:
      - name: checkout
        uses: actions/[email protected]
        with:
          fetch-depth: 2

      - name: autosense
        id: autosense                 # Ensure to set an 'id' field here or it might silently fail!
        run: sh autosense.sh          # Does something like: echo "::set-output name=runner::my-special-runner-label"

  build:
    needs: [init]                     # Mandatory dependency to ensure outputs are set
    runs-on: ${{ needs.init.outputs.runner }}
    run: ...

Conditional steps

Job and steps can be run conditionally by specifying if: clauses

job1:
  steps:
  - id: step1
    if: ${{ github.ref == 'main')

Note for boolean inputs always check on string 'true'!

 if: ${{ github.event.inputs[deployflag] == 'true' }}

There are additional string functions that can be used

 if: ${{ startsWith(matrix.stage, 'production-') == true }}
 if: ${{ endsWith(name, '_suffix') == false }}
 if: ${{ contains(mode, 'debug') != true }}

For ternary operator use <condition> && <outputTrue> || <outputFalse> syntax:

${{ startsWith(github.ref, 'feature') && 'this is a feature branch' || 'another branch' }}

To always run a step even if a previous step did fail

if: always()

Dynamic and customizable matrix job

Due to missing global variable support it is actually challenge to properly default a customizable matrix (i.e. when you allow it to be changed by workflow_dispatch input). Note that this is extra-complicated as you cannot access the default value of workflow dispatch events when you are in another event type. So the default matrix value has to be stored somewhere else and to be merged with any possible input. The only available global place for up to GHE 3.4 is a global env variable string.

name: testsuite

on:
  # The entire testsuite runs at night
  schedule:
  - cron: '0 2 * * *'
  
  workflow_dispatch:
    inputs:
      # Note: as we have a large nr of tests we want our users to be able
      # to only run some of them using this input
      testcases:
        description: Test cases to run (empty for all)
        type: string
        default: ""

env:
  testcases: tc1,tc2,tc3
        
jobs:
  # matrix preparation with extra step and fromJson() input as described by
  # https://stackoverflow.com/questions/65384420/how-to-make-a-github-action-matrix-element-conditional
  prepare:
    runs-on: self-hosts
    outputs:
      matrix: ${{ steps.set-matrix.outputs.matrix }}
    steps:
      - uses: actions/[email protected]
      - id: set-matrix
        run: |
          list="${{ github.event.inputs.testcases }}"
          list="${list:-${{ env.testcases }}}"
          echo '::set-output name=matrix::'"$(echo "$list" | jq -s --raw-input 'split("\n")[0] | split(",")' | tr "\n" " ")"

  testcase:  
    needs: prepare
    strategy:
      fail-fast: false
      matrix:
        name: ${{ fromJson(needs.prepare.outputs.matrix ) }}

    runs-on: self-hosted
    steps:
      [...]

Different working directory

For a single step:

jobs:
    job1:
       steps:
          - name: step1
            run: pwd
            working-directory: /etc

For all ‘run’ steps global default:

jobs:
  job1:
    defaults:
      run:
        working-directory: mysubdir

Inspecting all workflow_dispatch inputs

Use $ which will expand into a valid JSON map of all inputs and their values.

Job reports

To create a per job report from Markdown input.

cat <<EOT >>${GITHUB_STEP_SUMMARY}

## Report 

- info 1 [some link](https://example.com)
- info 2

EOT

You can also run it in each step to accumulate a report.

Manage Runners

Which jobs are run how often?

Print a pipeline histogram for the specific runner:

grep "value.:[[:space:]]*..github/workflows" /opt/app-root/src/_diag/Worker_*.log |\
sed 's/.*workflows.//;s/"//' | sort | uniq -c | sort -nr

Delete runners not alive anymore

This is useful with dynamically registered runners, that are not guaranteed to unregister (i.e. k8s pods)

  • List runners via /orgs/<org name>/actions/runners
  • Foreach runner with status == offline
    • Perform HTTP DELETE on /orgs/<org name>/actions/runner/<runner id>