Blog · November 17, 2026 · 12happy team
How to monitor your website's uptime for free, step by step
Most people find out their website is down from someone else: a customer, a client, a colleague who “just tried it”. By then it has usually been down for a while, and those first unnoticed minutes are part of what an outage costs you.
Uptime monitoring fixes that. A service outside your site requests it every few minutes, the way a visitor would, and tells you when the answer is wrong. It takes about ten minutes to set up, and for most sites it costs nothing.
This guide starts with checking your site right now, then sets up monitoring step by step.
Check whether your site is up right now
If you just want to know whether your site answers at this moment, run this in a terminal:
curl -sS -o /dev/null -w '%{http_code} in %{time_total}s\n' https://yoursite.com
200 in 0.41s means the page answered and how long it took. A 5xx code means the server answered
with an error. An error from curl itself, such as “Could not resolve host” or a timeout, means it
didn’t answer at all.
Two things to keep in mind:
- It’s one request, from one place. If it fails, try from another network, such as your phone with Wi-Fi off, before you conclude the site is down for everyone.
- It only tells you about now. An outage at 3 a.m. that’s over by 3:20 leaves no trace unless something was checking at the time. That’s what monitoring is for.
Step 1: decide what to watch
Start with what you’d actually act on. For most sites that’s:
- The home page. If it doesn’t load, nothing else matters.
- The page that earns money or leads: the checkout, the booking page, the contact form. It can break while the home page looks fine.
- Anything your product calls, such as an API. If it has a health endpoint (often
/healthor/status), watch that. - Servers that aren’t websites: a mail server or a database port can be watched with a TCP check, and a server’s network with ping.
Three to five monitors cover most small sites. You can always add more.
Step 2: choose a monitoring service
Free plans differ more than they look. Before you choose one, check:
- How often it checks. Five minutes is common on free plans. An outage shorter than the interval can fall between two checks.
- How many monitors you get, and whether the free plan allows a business site.
- Whether it confirms a failure from a second location before alerting. Without that, one bad network path wakes you up for nothing.
- Where alerts can go. Email only, or also Slack, Telegram and your phone?
- Whether it reads the page. A keyword check catches pages that load but show the wrong thing.
- Whether there’s a status page, so your users can see what’s going on without asking.
The steps below use 12happy’s free plan: 50 monitors checked every 5 minutes, HTTP, keyword, ping and TCP checks, every failure confirmed from a second location, alerts by email, Slack, Telegram and the mobile app, and one status page. The same steps work, with different labels, in most monitoring tools.
Step 3: add your first monitor
- Sign up and open the console.
- Choose Add monitor.
- Under What should we watch?, enter your address, for example
https://yoursite.com. - Under Check type, choose one:
- HTTP checks that the page answers with 200 OK, after following any redirects. Start here.
- Keyword also checks that a piece of text you choose is on the page. Use text near the bottom of the page that only appears when it’s working, such as a phrase from your footer.
- Leave How often at the fastest your plan allows. On the free plan that’s every 5 minutes.
- Under Tell these channels, tick where alerts should go. If you haven’t set any up yet, do step 4 and come back to it.
Give the monitor a name you’ll recognise in an alert at 3 a.m., such as “Shop: checkout” rather than the bare address.
Step 4: send alerts where you’ll see them
A monitor that alerts an inbox nobody reads isn’t monitoring. Pick the place you’d notice within minutes.
- Open Alert channels and choose Add channel.
- Choose the type:
- Email: fine as a second channel, easy to miss as the only one.
- Slack or Telegram: good for teams, and for anyone who lives in those apps.
- Push: the 12happy app for iPhone or Android. The most reliable way to reach one person.
- Choose Send a test, and check that the test arrives.
- Go back to your monitors and make sure each one is sending to the channel.
Two channels are better than one. If Slack is having a bad day, your phone still buzzes.
Step 5: test it with a real failure
A test alert proves the channel works. It doesn’t prove the monitor notices a failure. To check the whole chain:
- Add a temporary monitor for an address that doesn’t exist on your site, such as
https://yoursite.com/monitoring-test-404, with the HTTP check type. - Wait. The page answers 404, the failure is confirmed from a second location, and the alert arrives. On a 5-minute interval, expect it within a few minutes.
- Delete the test monitor.
If no alert came, check that the monitor is routed to a channel, and that the channel isn’t paused.
Step 6 (optional): add a status page
When your site is down, people want to know whether you know. A status page answers that without an inbox full of “is it just me?”.
In the console, open Status pages, choose New page and pick the monitors to show. The page shows each one’s current state and its uptime over the past 90 days, and updates itself from your monitors. On the free plan it lives on a 12happy address. See what the free status page includes. To show its state in your site’s footer or a README, add a live status badge.
When an alert arrives
You’ll want a short routine for the moment it happens:
- Open the site from another network, such as your phone’s mobile data. Does it fail for you too?
- Look at the error in the alert. A timeout points to the server or the host. A 5xx points to the application. A certificate error means it expired or was replaced.
- Check what changed. A deploy, a plugin update or a DNS change in the last hour is the usual suspect.
- Check your host’s status page. Sometimes it isn’t you. For Cloudflare, Vercel, WP Engine and other common services, our live status pages show our own checks next to the vendor’s.
- Tell people. Acknowledge the incident in the console so your team knows someone is on it, and post an update on your status page if you have one.
The monitor closes the incident by itself when the site recovers, and you get a second alert saying so, with how long it was down.
Is every 5 minutes enough?
For a personal site, a blog or a portfolio, yes. For a shop or a service people pay for, a 4-minute outage can fall between two checks and never show up. If you promise an uptime figure to a client, your check interval has to be able to prove it: the nines of availability explains why, and the SLA calculator shows what each target allows. Paid plans check every 30 seconds (Pro) or 15 seconds (Agency).
Running WordPress? Some of its failures still answer 200. This guide covers the checks that catch them.