Start with the customer journey that matters.
Monitor your public website and a safe read-only endpoint for the service behind it. If a scheduled backup matters, add a heartbeat sent after that job completes.
Choose one responsible recipient and an escalation process you can maintain. Verify delivery with a controlled rehearsal and revisit alerts after changes to your website or mail provider.
Begin with a few useful checks. More monitors create more observations and storage; they do not automatically create better incident response.