How to Get a Website Down Alert by Email
There are a few ways to get a website down alert by email: a free uptime monitor like UptimeRobot, the downtime alert your host or platform already offers, a checker you run yourself such as Uptime Kuma, a paid service that rings a phone, or one sentence to MailMeWhen, with no account and no dashboard.
Most people don't find out from a dashboard. A customer writes to ask if you're still open. Or you open your own shop on a Sunday evening and get a blank page. By then it has been down for hours, and every order and every sign-up in that time is gone.
Ways to get a website down alert by email
A website down alert in one sentence
Write down the address and what 'working' means to you: 'mail me if my shop stops loading'. A real person sets it up. No account, no dashboard, nothing for you to keep configured. We check the site every minute, and as long as it answers, you hear nothing.
When it stops answering, we try again over a completely different network route. We only write when both fail, because one failed route can mean our line is broken, not your site. The letter reaches you two or three minutes after the site went quiet. It says in plain words what we saw: a request that timed out, a 500 error from the server, or a domain name that no longer resolves. If the name can still be looked up, the letter says whether it still points at a server, which tells you whether to call your host or your registrar. When the site comes back, a second letter says how long it was down in total.
If you like, a short note every week tells you the site has been fine and we're still watching. With any monitor, silence can mean nothing happened or the monitor itself died, and that note tells the two apart. The first 7 days are free with no payment details; after that it is $5/month for that one site, and it stops by itself if you don't subscribe. Reply once to stop.
Good to know about downtime alerts
Checking only the home page can miss the part that's actually broken. A shop whose checkout has failed often still shows a perfect home page. An app whose backend is dead still draws its own frame. So name a word that has to be on the page, like 'Add to basket' or 'Log in'. Then a page that loads but no longer works counts as down too.
False alarms are why people stop reading alerts. One failed check is not an outage. A restart, a dropped packet or one bad minute on the monitor's own line looks exactly like a dead server from outside. Whatever you use should check again before it writes, and ideally from a different network.
Knowing is not fixing. No alert keeps a site online. What it gives you is the time between the site going quiet and a customer noticing. Decide in advance who gets the letter and what they do about it. And if an hour of downtime is expensive, pay for something that rings a phone at three in the morning.
Try it — hand us this sentence:
Watch my shop at [paste the link] and mail me if it stops loading, and again when it's back up. The page should say 'Add to basket'.
Common questions
Isn't there a free website down alert?
Yes, and for many sites it's the right answer. UptimeRobot's free plan covers 50 addresses, Jetpack watches a WordPress site for free, and Uptime Kuma is free if you have a server to run it on. They check every five minutes and expect you to keep an account and a dashboard. MailMeWhen is the other trade: one sentence, nothing to set up, a check every minute, and letters in plain words. If you're comfortable with a dashboard, use the free one.
How quickly will I know that my website is down?
We check every minute and confirm the failure over a second, independent route before writing, so the letter usually reaches you two to three minutes after the site stopped answering. Within 24 hours of your request, a first letter tells you what is being watched, how often, and what the first check saw. There is no interval to choose and no dashboard to keep.
The monitor says my site is down, but it loads fine for me. Which is right?
Both can be right. A site can be up for one network and unreachable from another: a blocked range, a broken route, a regional problem at the CDN, or a DNS change that some resolvers have and others don't. That's why one failed check should never become a letter. We call it down only when two independent routes fail. When they disagree, we stay quiet and look again rather than tell you something we can't prove.