Website Monitoring Basics: What to Check and How Often

Start with availability

Check public endpoints from more than one region and verify expected status, content, and certificate validity. A 200 response can still be a broken page, so include a small content assertion on critical journeys.

Add experience signals

Track response latency, error rate, and representative browser journeys. Use synthetic checks for known paths and real-user measurements for population trends. Keep bots and internal health traffic out of business conversion reports.

Define alert conditions

Alert only when someone can take a named action. Require repeated failure for noisy networks, group related symptoms, and include evidence in the notification. Separate warning, incident, and reporting thresholds.

• Critical journeys listed
• Two-region availability check
• Certificate expiry warning
• Owner and runbook linked

Review the monitor itself

Test alerts, escalation, and recovery notices. Remove checks that never lead to action. After incidents, update the monitor to detect the earliest useful symptom instead of adding a dozen overlapping alerts.