How UptimeRobot prevents false alerts.
When a check fails, a second data center in the same region checks again before we alert you, on every plan, Free included. With multi-location monitoring, one monitor checks from up to four regions, so you also catch outages that only hit one part of the world.
No credit card required. Nothing to install.
7.5M+
false downs stopped every month
That's around 250,000 monitors a day that would have gone red while the site was up.
2.4B
checks a day
Each one that fails is checked again from a second data center before anyone gets an alert.
Why false alerts happen.
A false alert is a down notification for a site that was up the whole time. They often start between the checker and your server. A connection drops, a DNS lookup times out, or one network route is congested for a few seconds. Sometimes the site itself stalls for a few seconds during a deploy or restart.
A monitor that trusts that single check pages you at 3 a.m. for nothing. After a few of those, people start muting alerts, and the next real outage gets missed.
How the re-checks work.
A monitor goes down only after re-checks from other data centers in the same region confirm the failure. Until then, nobody gets an alert.
Catch the outages only part of the world sees.
Check one monitor from up to four regions. When your site goes down in one of them, you get an alert even if your own region looks fine.
Why not just self-host an open-source monitor?
Open-source tools like Uptime Kuma are free to download, but by default they check from the one server you install them on. Retries, if you set them up, run from that same server, so when its connection drops, every monitor on it can go red at once while your sites are fine. And if the server goes down, your monitoring goes down with it.
Self-hosted open source | UptimeRobot | |
|---|---|---|
Where checks run | Servers you run yourself | Our data centers |
After a failed check | Retries, if you set them up, run from your own servers (none by default in Uptime Kuma 2) | A second data center in the same region checks again before we alert you |
Checks from other regions | Only from places where you set up and run a server | Up to four regions per monitor |
Who keeps the checker running | You: the server, updates, backups Running self-hosted monitoring comes with a cost. Work out yours with the self-hosting cost calculator. | UptimeRobot |
Nothing to install, nothing to keep running.
Your checks run from our data centers, and a failed check gets a second look from another one before we alert you.
50 monitors free. No credit card required.
What your plan includes.
The re-check runs on every plan, Free included. Paid plans add location-specific monitoring, so you can check one monitor from up to four regions and catch outages that only affect part of the world.
Free | Solo | Team | Scale | |
|---|---|---|---|---|
Re-check from a second data center | ✓ | ✓ | ✓ | ✓ |
Monitoring interval | 5 minutes | 60 seconds | 30 seconds | 15 seconds |
Location-specific monitoring (up to four regions) | Not included | ✓ | ✓ | ✓ |
Trusted by 3.5M+ developers and teams.
More ways to get alerts you can trust.
Delay alerts.
Set a monitor to alert only when it has been down for a set number of minutes, and repeat the alert every few minutes while it stays down.
Check for the right content.
A keyword monitor checks that your page shows the text it should, so a page that loads but shows an error still counts as down.
Schedule maintenance windows.
Pause monitoring during planned work, so expected downtime doesn't trigger alerts or count against your uptime.
Frequently asked questions.
A false alert, or false positive, is a down notification for a service that was actually up. It usually comes from a problem between the monitoring checker and your server, such as a dropped connection, a DNS timeout or a congested network route. UptimeRobot re-checks a failed check from other data centers in the same region before alerting, so you only get alerts for failures the re-checks confirm.
When a check fails, UptimeRobot re-checks it within seconds from a different, randomly chosen data center in the same region. The monitor goes down and you get an alert only when the re-checks confirm the failure. This covers timeouts, network and DNS errors, error status codes and keyword or assertion mismatches. Every month, the re-check stops over 7.5 million false downs, around 250,000 monitors a day.
Yes. The re-check from a second data center runs on every plan, Free included. Free monitors check every 5 minutes. Paid plans check more often (60 seconds on Solo, 30 seconds on Team, 15 seconds on Scale) and add location-specific monitoring from up to four regions.
Coverage. One monitor checks from up to four regions, so you catch outages that only part of the world sees, even when your own region looks fine.
Common causes are a cached error page that our checker still receives, or a deploy or restart that keeps your site from responding while we re-check. Our guide to debugging a monitor that shows as down walks through the possible causes and fixes.
Only if you set retries up, and they run from your own server, so a problem there can still look like an outage. In Uptime Kuma 2, for example, new monitors have no retries by default, so the first failed check can trigger an alert.
Double-checked alerts for free, 50 monitors included!
Join 3.5M+ users who rely on UptimeRobot to confirm failures before sending alerts.
No credit card required. Nothing to install.