A surprising number of businesses discover outages through a customer email, hours after the fact. Basic monitoring costs very little and converts that into a notification within minutes, which is usually the difference between a brief interruption and a lost day.

Check a journey, not a heartbeat

A homepage returning 200 tells you the web server is alive. It does not tell you whether the database is reachable, whether checkout works, or whether the contact form is delivering. Add a health endpoint that verifies real dependencies, and monitor at least one transactional path end to end.

Monitor the things that expire

  • TLS certificate expiry, with alerts well before the date.
  • Domain registration expiry — expired domains cause complete, embarrassing outages.
  • API keys and credentials with expiry dates.
  • Scheduled job completion, so a job that silently stops is detected.

Watch performance, not only availability

Sites degrade before they fail. Track response time and alert on sustained increases, and watch error rates as well as hard failures. A rising rate of 500s on one endpoint is the early warning that precedes the outage everyone notices.

Route alerts to somewhere real

An alert into an unmonitored inbox is not an alert. Decide which notifications justify waking someone and which can wait until morning, and route them differently. Test the path occasionally — alerting configurations break silently, and the failure is only discovered during the incident they were meant to catch.

Keep the alert count low

Alert fatigue is the predictable end state of monitoring everything. Start with four or five high-signal alerts: site unreachable, error rate spike, certificate expiring, backup failed, and a critical journey failing. Add more only when a specific incident proves the need.

Write the runbook alongside

Each alert should link to a short note: what it means, what to check first, and who to contact. This transforms a 2am alert from an investigation into a procedure, and it is what makes monitoring useful to somebody other than the person who configured it.