Based on feedback from new user who was pointed to this place, we should update the document to show:
- How to create server-side monitoring configuration
- How to write alarm criteria, especially make it clear that only the metric, not the status, can trigger alerts.