Uptime Monitoring: Frequencies, Locations and False Alarms

Every monitoring setup answers three questions badly or well: how often do you check, from where, and what counts as "down". Get them wrong in either direction and you either miss real outages or learn to ignore your own alerts.

How often is often enough?

A 10-minute interval is fine for a brochure site. For an API or an e-commerce checkout, every minute or less: at €9.99/month plans allow 1-minute checks; you trade a little noise for a maximum blind window you can actually budget for.

One location lies

A monitor in a single data center shares fate with that provider: a BGP hiccup there becomes a "site down" page for you. Multi-location monitoring verifies an outage from at least two independent regions before firing the alert — this single change eliminates most false positives.

What counts as down?

• Non-2xx/3xx status codes your app never intentionally returns
• Timeouts and TLS handshake failures, not just connection refused
• Content checks: a keyword or JSON field on the page, not just a 200

SiteFlux implements all three natively: multi-location verification, response-time history, and keyword validation per monitor, with escalation channels from email up to automated phone calls.