Why we check every minute, not every five

The industry default is five minutes. Almost every uptime tool you can name ships with that cadence on their paid tier. UptimeRobot, Pingdom, StatusCake — all built around the same quiet assumption: that a site being down for four minutes and fifty-nine seconds doesn't matter.

It does.

The math nobody wants to do

If your checks run every five minutes, the average time-to-detect for a complete outage is 150 seconds. The worst case is 299 seconds, the last check having landed one second before the outage began.

That's five minutes you don't know. Five minutes your users are refreshing the page, opening tickets, writing angry tweets, and googling your competitors. Five minutes in which the person who needs to notice is still eating lunch.

Bring the interval down to sixty seconds and those numbers become 30 seconds average and 59 worst-case. You aren't "monitoring more frequently." You are shortening the blast radius of every outage by five times.

Why the five-minute default persists

The honest answer is cost. A monitor checking every five minutes generates 288 HTTP requests per day. A monitor checking every minute generates 1,440 — five times the infrastructure, five times the outbound bandwidth, five times the log volume.

That is real money when you are running legacy infrastructure at scale. It is rounding error when you are running on modern serverless edges.

What sixty seconds actually gets you

• A single flapping check doesn't fire an alert. We wait for two consecutive failures before sending mail. At one-minute cadence that is at most two minutes of confirmation. At five-minute cadence it is ten.
• Incidents have a believable start time. If your status page says "down since 14:22" and the real answer was 14:21, readers stop reading fine print. They trust the clock.
• Recovery is fast. The moment the next check succeeds, the incident closes and the all-clear alert lands. You don't spend five minutes convinced the site is still on fire.

Every one of those improvements is worth more than the dollars the extra checks cost us.

The real reason

The real reason we ship at sixty seconds is that we believe infrastructure monitoring is a category where the user — not the vendor — pays the cost of slow detection.

We didn't want to be the vendor that makes that tradeoff quietly on your behalf.

A monitoring tool that doesn't check often enough to notice is a monitoring tool that doesn't notice.

Five-minute monitoring is a polite fiction. Sixty-second monitoring is what you actually wanted when you signed up.