Uptime checkers that test from one place have a known weakness. A hiccup between the checker and the site looks exactly like an outage. Get a few of those alerts at night and people start ignoring them, which defeats the point of monitoring.

I built SitePulse, a small uptime monitor for freelancers and agencies, around a single rule. It should only alert when it is fairly sure the site is down.

How it decides a site is down

A check fails when the site doesn't answer within 10 seconds, or answers with an error status. One failure isn't enough. It has to fail twice in a row.

Then it asks for a second opinion. The same site is checked from Europe and from Asia. If either region can still reach it, the trouble was on the checker's side of the internet and nobody gets an alert. Only when those checks fail too does the site count as down.

A redirect counts as up, because the server answered.

The trade-off

This rule has a cost, and I'd rather say it plainly. If a site is unreachable from one part of the world but fine from the others, SitePulse won't call it down. I chose this side on purpose: for most small sites, a false alarm at night costs more trust than a missed regional blip. If you need alerts per country, this isn't the right tool.

Waiting for the second opinion also means the alert arrives a little later than the first failed check.

A site can answer and still be broken

A status code doesn't tell you everything. A shop can return 200 while its product page lost the "Add to cart" button after a bad deploy. SitePulse lets you add text the page must contain. If that text disappears, the site counts as down, even while the server keeps answering.

Expired domains

Each domain's expiry date is looked up every day, and you get an email 30, 7 and 1 days before it runs out. Under the hood that's RDAP where the registry supports it, with a WHOIS fallback where it doesn't. Some registries don't publish the date at all. Then SitePulse shows that the date isn't published instead of guessing one.

For people who look after other people's sites

These are the parts I built for agencies and freelancers:

  • A private status page per client: one link, no login, showing only that client's sites, uptime and outages. You can turn the link off any time.
  • A report on the 1st of each month with last month's uptime, outages and response times, with your agency's name at the top. When the client replies, the reply comes to you.
  • Alerts in Slack, Discord, Microsoft Teams or your own signed webhook, with times shown in each reader's time zone.

They're on the Pro plan, which isn't open for upgrades yet.

How it's built

SitePulse runs on Cloudflare. Workers serve the app and run the checker from a cron trigger every minute, and the data lives in D1. The second opinions come from Durable Objects placed in Europe and Asia with location hints. When I tested it, those checks really did run from those regions.

Try it

The free plan covers 5 sites checked every 5 minutes, with email alerts, the second opinion from Europe and Asia, and "page must contain" checks. You sign in with a link sent to your email, so there's no password and no card.

I built SitePulse, so I'm biased. If you try it, I'd like to hear where it falls short: https://sitepulse.buymazing.in