The problem
Most small sites fail at the wrong time. A client's shop stops taking orders on a Friday evening. A certificate expires over a bank holiday. A server runs out of disk space at three in the morning. You usually find out when someone emails you, and sometimes not until they stop replying.
If you run your own web app or look after client sites, you are the person who gets the message when something breaks. There is no operations team behind you. The gap that matters is the time between the site going down and you knowing about it.
A website uptime monitor with alerts closes that gap. It checks your site on a schedule and tells you when the site stops answering the way you expect. This guide explains how that works, what to alert on, and how to set one up without much fuss.
How it works
A monitor runs a simple loop. At a set interval, the service requests an address from outside your own network. It records whether a response arrived, how long it took, and what status code came back. If the answer is wrong, the monitor is marked as down and an alert is sent.
There are four decisions to make for each monitor.
First, the address. Pick a page that exercises the thing you care about. The homepage is fine for a brochure site. For a web app, a health page that reads from the database tells you more than a static file does.
Second, what counts as failure. A timeout, a refused connection, or a 5xx status code is an obvious failure. A page that returns a 200 but shows an error message is a failure too. Checking for a short phrase that should always appear on the page catches this kind of breakage.
Third, the interval. Shorter intervals catch problems sooner but make more requests. For most client sites, a check every few minutes is enough.
Fourth, the threshold. One failed check can be a blip, such as a dropped connection between two networks. Many people wait for two or three failures in a row before they alert, so they are not woken for a single missed response. Alerts should also go to the right people. On AperioFlow, alert emails go only to people in your organisation.
There is a second kind of check that is easy to forget. A website monitor asks whether your site answers. A heartbeat monitor waits for your own job to report in. Your nightly backup or cron job sends a request to a unique address when it finishes. If that request does not arrive on time, the heartbeat monitor alerts you. This is useful for backups, because a backup that fails quietly is the one you discover when you need it. A website check cannot tell you that, but a heartbeat can.
When you are choosing a tool, check where your data is kept and whether you can take it with you. AperioFlow is hosted in France and run from the UK, and you can download your data.
Getting started
You do not need to monitor everything on day one. Start with the parts that would hurt most if they broke, and add more once you trust the setup.
- Write a short list. For each site, note the homepage, any page that takes money or sign-ins, the API endpoint your app depends on, and any scheduled job that must run.
- Add one monitor for each item on the list. Use the full address, including https, so the check runs the same path a visitor would.
- Set the expected result. At minimum, choose the status code you expect. If you can, add a phrase that only appears when the page is working correctly, such as the name of your product.
- Set the failure threshold to two or three consecutive failures, then adjust after a week. If you are getting alerts for brief blips, raise it. If a real outage goes unnoticed for too long, lower it.
- Decide who gets the alerts. Send them to the person who can fix the problem, and to a second person if you have one. Keep the list inside your organisation.
- Test the alert before you rely on it. Point one monitor at a deliberately wrong address, confirm the alert arrives, then change the address back.
- Set up a status page if your customers need to know what is happening. A status page gives them one place to check, so they do not have to email you to ask whether the site is up.
- When something does break, post an incident update early. Say what is affected, what you know so far, and when you will post the next update. Short, factual updates do more to keep trust than a long explanation written later.
AperioFlow has a free plan with 10 monitors, so you can set up checks for a small number of sites without paying anything. The Team plan is 29 pounds a month flat for up to 10 people, as shown on the pricing page. Tweed Tyne runs its own work on AperioFlow, which is one reason we keep the monitoring and alerting parts simple.
If you have a site that matters and nothing watching it yet, start with the homepage and one page that takes money or sign-ins. Start free at https://app.aperioflow.co.uk/#/signup and add the rest once the first alerts are working the way you want.