Most small teams find out a site is down from a customer. Someone emails, posts in a chat, or leaves a review. By then the site may have been broken for an hour or more. Uptime monitoring exists to shorten that gap.
The problem
If you run a web app or a client site, you are often the only person who gets the message when something breaks. There may be no operations team, no on-call rota, and no dashboard that someone watches over breakfast. You are the dashboard.
That works until you are on a train, asleep, or deep in a task for another client. A payment page returns an error for twenty minutes. A certificate expires on a Sunday. A nightly backup stops running and nobody notices for a month. None of these failures are dramatic. They are quiet, and quiet failures cost money and trust.
Small teams tend to have three specific problems:
- Nobody owns the site full time, so checks are ad hoc or forgotten.
- Alerts land in one inbox, where they get lost among invoices and newsletters.
- When something does break, customers ask what is happening, and there is no single place to answer.
Uptime monitoring tackles the first problem. Alerts and a status page help with the second and third. The aim is not perfection. The aim is to know sooner and to say something useful while you fix it.
How it works
An uptime monitor is simple in principle. It checks a URL on a schedule, usually every few minutes. If the page returns an error, times out, or returns something unexpected, the monitor records a failure and sends an alert once the failure is confirmed. When the site recovers, it records that too.
Good monitoring depends more on the alert rules than on the check itself. Decide what counts as down before you set anything up:
- A page that returns a server error (a 5xx response) is almost always worth an alert.
- A slow response may be worth a warning rather than an alert that wakes someone up.
- A login page or checkout that loads but fails a key step needs a check that goes beyond the status code, if your tool supports it.
- Certificate expiry is worth an alert days in advance, not on the day.
Keep alerts short and specific. 'Checkout returning 500 for 3 minutes' is useful. 'Monitor alert' is not. Alert emails go only to people in your organisation, so make sure at least two people receive them. That way one person's holiday does not leave the site unwatched.
Heartbeat monitors cover a different case. A heartbeat is a ping your code sends when a job finishes. Cron jobs and backups are the usual examples. If the ping does not arrive on time, the monitor raises an alert. This catches the job that silently stopped, which a normal URL check would never see. A nightly backup that has not reported in for 26 hours is a failure worth knowing about the same morning.
A status page sits on top of the monitors. It shows the state of each service to your customers, and it is where you post incident updates. Customers who can see what you already know tend to ask fewer questions.
Getting started
You do not need a long project to begin. A workable setup takes an afternoon:
- List the services that matter. This is usually your main site, your app's login, your payment or booking step, and any API your customers rely on.
- Add one monitor per service. Start with the public URL, then add a check for the key step if the first one is too shallow.
- Set an alert rule for each. Decide how long a failure must last before it alerts, so a single blip does not wake anyone.
- Add a heartbeat for each scheduled job, starting with backups, because a failed backup is the most expensive silent failure.
- Agree who receives alerts and who answers them. Write it down in one line.
- Create a status page and link to it from your site footer, so customers know where to look.
- Test it. Pause a job or break something on a staging copy, and confirm the alert arrives.
AperioFlow brings this into one place. You can watch uptime, set up heartbeats for cron jobs and backups, catch errors, run a status page, and post incident updates, alongside your task planning. The Free plan includes 10 monitors, which is enough to cover a small set of sites while you work out what needs watching. The Team plan is £29 a month flat for up to 10 people, as listed on the pricing page. Hosted in France. Run from the UK. You can download your data at any time.
Start free at https://app.aperioflow.co.uk/#/signup and set up your first monitor this afternoon.