Uptime monitoring involves automated checks — usually simple HTTP requests — sent to a site at regular intervals (e.g., every minute). If the site fails to respond or returns an error status code, an alert is triggered so the operator can investigate immediately.

Why it matters: Even brief undetected downtime costs revenue, search ranking, and user trust. Without monitoring, outages are reported by customers, not caught proactively. A 99.9% uptime SLA still permits ~8.7 hours of downtime per year.

Quick tips:

Monitor from multiple geographic locations to distinguish a regional outage from a global one.
Check both HTTP 200 responses and meaningful content (a "maintenance mode" page still returns 200).
Set alert thresholds carefully — instant alerts on a single failure lead to alert fatigue; 2-of-3 confirmations reduce false positives.