Automatically create GitHub issue when problems are detected

[PGijsbers]

When the workflow runs, if a problem is detected:

• Detect if there already is an open issue that reports this problem.
• If so, do nothing. If not, open an issue and ping core dev(s).
• In any case, embed a link to the related issue in the service card.

I would refrain from automatically closing issues when the workflow no longer detects issues, but perhaps a comment could be posted.

[axif0]

We can add a scheduled monitor workflow to the status repo. Every 10 minutes (and on manual dispatch) it runs generate_page.py , captures any failure output, and using actions/github-script@v7 automatically opens a GitHub issue with the log attached so we have a breadcrumb trail when checks break. The workflow then re-fails the run so incidents still show up red in Actions.

name: "Monitor Status Checks"

on:
  schedule:
    - cron: "*/10 * * * *"
  workflow_dispatch:

permissions:
  contents: read
  issues: write

jobs:
  monitor:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: actions/setup-python@v5
        with:
          python-version: "3.10"

      - name: Run status checks
        id: checks
        continue-on-error: true
        run: |
          set -o pipefail
          python generate_page.py 2>&1 | tee monitor.log

      - name: Create incident issue
        if: steps.checks.outcome == 'failure'
        uses: actions/github-script@v7
        with:
          github-token: ${{ secrets.GITHUB_TOKEN }}
          script: |
            const fs = require('fs');
            const body = fs.readFileSync('monitor.log', 'utf8');
            const timestamp = new Date().toISOString();
            await github.rest.issues.create({
              owner: context.repo.owner,
              repo: context.repo.repo,
              title: `Status monitor failure (${timestamp})`,
              body: `generate_page.py failed during the scheduled monitor run.\n\n\`\`\`\n${body}\n\`\`\``
            });

      - name: Fail job when checks failed
        if: steps.checks.outcome == 'failure'
        run: exit 1

@PGijsbers is this all make sense ?

[PGijsbers]

Sorry for the late reply. Yes that makes sense! On second thought it would also be good if there already is an open issue to just comment on it. So we can avoid a new issue being created every 10 mins in case of longer downtime.

[axif0]

it would also be good if there already is an open issue

Can you help create the initial issue so I can update the workflow to comment on it instead of opening a new one each time?
@PGijsbers