GitHub Actions步骤自取消问题解决及通知方案咨询

问题场景 #

现有一个GitHub Actions Workflow步骤，仅当分支名称以/feature结尾时执行，包含依赖安装、Poetry配置及dbt相关Make命令执行操作，目前该步骤被Workflow自动取消，需排查原因并实现异常捕获通知。

对应的步骤代码如下：

- name: run
  if: endsWith(github.ref, '/feature')
  run: |
    echo "installing dependencies"
    pip3 install --upgrade pip
    pip3 install poetry
    cd job-extract/trans/project-v2
    poetry install
    echo "executing dbt"
    make source_dbt
    make run_catalog_server
    make run_scheduled_refresh

问题排查思路 #

常见取消原因及对应排查动作 #

1. 执行超时：GitHub Actions默认单步骤/整个Job有6小时超时限制，若某个Make命令执行时间过长会被自动终止
   • 排查：在Workflow的Job或步骤中显式设置timeout-minutes（如timeout-minutes: 120）测试是否恢复；拆分长步骤为多个独立步骤，定位具体超时环节
2. 资源耗尽：Runner的内存/CPU不足，导致进程被系统终止
   • 排查：在步骤中添加资源监控命令（如free -h、top -b -n 1）输出资源使用情况；尝试使用更大规格的Runner（如GitHub托管的ubuntu-latest换成更大实例，或自托管Runner）
3. 命令阻塞/无退出：若run_catalog_server这类Make目标是启动常驻后台服务，会导致脚本一直阻塞，Workflow判定任务停滞而取消
   • 排查：检查Makefile中对应目标的实现，确认是否有后台启动逻辑；若为临时服务，需添加前台运行或自动退出的逻辑
4. 未捕获的错误：某个命令执行失败但脚本未中断（如poetry install失败但继续执行后续命令），导致Workflow异常处理触发取消
   • 排查：在脚本开头添加set -e（遇到错误立即退出），并为每个关键命令添加日志输出，定位失败环节：

     run: |
       set -e
       echo "Upgrading pip..."
       pip3 install --upgrade pip
       echo "Installing poetry..."
       pip3 install poetry
       cd job-extract/trans/project-v2
       echo "Running poetry install..."
       poetry install
       echo "Starting dbt commands..."
       make source_dbt
       echo "Completed source_dbt"
       make run_catalog_server
       echo "Completed run_catalog_server"
       make run_scheduled_refresh
     

异常通知实现 #

通过GitHub Actions的条件判断failure()或cancelled()捕获失败/取消状态，结合第三方动作实现通知：

示例1：Slack通知 #

- name: Send Slack notification on failure/cancellation
  if: failure() || cancelled()
  uses: slackapi/slack-github-action@v1.24.0
  with:
    slack-message: "⚠️ Workflow步骤执行异常：分支${{ github.ref_name }}，Job: ${{ github.job }}，运行ID: ${{ github.run_id }}"
  env:
    SLACK_WEBHOOK_URL: ${{ secrets.SLACK_WEBHOOK_URL }}

示例2：创建GitHub Issue #

- name: Create issue for failed workflow
  if: failure() || cancelled()
  uses: actions/github-script@v6
  with:
    script: |
      github.rest.issues.create({
        owner: context.repo.owner,
        repo: context.repo.repo,
        title: `Workflow异常：${github.ref_name}分支步骤取消/失败`,
        body: `Job名称：${github.job}\n运行链接：https://github.com/${context.repo.owner}/${context.repo.repo}/actions/runs/${github.run_id}`
      })

示例3：邮件通知 #

- name: Send email notification
  if: failure() || cancelled()
  uses: dawidd6/action-send-mail@v3
  with:
    server_address: smtp.gmail.com
    server_port: 465
    username: ${{ secrets.EMAIL_USERNAME }}
    password: ${{ secrets.EMAIL_PASSWORD }}
    subject: GitHub Actions Workflow异常通知
    body: |
      分支：${{ github.ref_name }}
      Job：${{ github.job }}
      状态：${{ job.status }}
      运行链接：https://github.com/${context.repo.owner}/${context.repo.repo}/actions/runs/${github.run_id}
    to: your-email@example.com
    from: GitHub Actions Bot

内容的提问来源于stack exchange，提问作者bionics parv