Why We Report a Range, Not an Exact Time
Our alerts say things like “down at least 6 minutes, at most 12.” Here's why that's not vagueness — it's the only honest answer any monitor can give.
An exact number would be a guess
We watch your site from the outside, the way your visitors experience it — checking at intervals, like a camera taking a photo every few minutes. An outage can begin or end between two photos, and whatever happens there is invisible to us. It's invisible to every other monitoring service too; only your own server logs know the exact second.
The difference is what's done with that uncertainty. Most tools pick one number inside the blind window and present it as fact. We won't dress a guess up as a measurement — so we tell you the two numbers we actually know: the span we directly observed down, and the widest the outage could have been. If our report ever disagrees with your server logs, the truth will be inside our range. We verify that promise against scheduled outages we run on our own infrastructure, every night.
The end of an outage is pinned tightly
The moment a check fails, we stop ambling and start sprinting: checks accelerate automatically (every 60 seconds on paid plans, every 5 minutes on Free) until your site recovers. That means the endof an outage — the part you usually care about most, “is it back?” — is always known within about a minute on paid plans. Only the start of an outage inherits your regular check interval, because the failure happened somewhere between two routine checks.
Why not just check every second?
Two reasons. First, every check is a real request that your server answers. At one-second intervals that's 86,400 extra hits a day — at some point the monitor becomes your problem instead of watching for it.
Second, it still wouldn't be “exact” — just a smaller range. The blind window between checks never disappears; it only shrinks. The honest lever isn't pretending the window away, it's choosing how small you want it:
- Free — uptime checked every 5 minutes; everything else daily.
- Starter — uptime every 5 minutes; other checks hourly.
- Pro — every 5 minutes.
- Business — every minute, so the range is rarely wider than a minute or two.
Every plan accelerates the instant something looks wrong — a tighter plan buys you a tighter start-of-outage bound, not a different kind of truth.
The short version
When you see “down at least 6m, at most 12m,” read it as: we verified 6 minutes of downtime with our own eyes, and it cannot have been longer than 12. That's everything an outside observer can honestly claim — and it's more information, not less, than a single made-up number.