Monitoring · updated 6 August 2026
Uptime monitoring for small businesses: what to set up and what to ignore
A pragmatic setup for a site with no ops team — what to monitor, how often, who to tell, and the upsells not worth paying for.
Most small businesses learn about downtime from a customer, hours after it started. Fixing that does not need a monitoring platform, an on-call rota or a per-seat subscription. It needs one check, one email address, and about ten minutes.
Monitor the page that matters, not the home page
The home page is usually the most heavily cached thing you own, which makes it the least likely to notice a failure. If your business depends on a booking form, a checkout or a login, point the monitor at a URL that actually exercises that path. A tiny health-check page that touches the database and returns "ok" is the ideal target: cheap to serve, and impossible to serve from cache when the thing behind it is broken.
Five minutes is the right interval
One-minute checking sounds better and rarely is. A single failed check is never enough to act on — every honest system waits for a second confirmation — so the practical difference between one-minute and five-minute monitoring is a few minutes of notice, in exchange for five times the load on your own server. Unless downtime costs you meaningful money per minute, five minutes with confirmation is right, and it is what our free monitoring does.
Alert the person who can act
An alert to an inbox nobody reads on a Sunday is decoration. Send it somewhere a human will see it, and if your site is run by an agency or a freelancer, send it to them as well as to you — you want the person with server access to know first. Keep the alert list short; alerts that go to everyone get muted by everyone.
Watch the certificate too
Expired TLS certificates cause more sudden total outages at small businesses than server failures do, because automated renewal breaks silently. A monitor catches it when the site goes down; better is to check the expiry date directly once a month with the SSL checker and act at 20 days remaining. Same for the domain itself — a lapsed renewal on an old card takes a business offline completely, and the WHOIS lookup shows the expiry date.
What to ignore
- SMS and phone alerts — worth paying for only if downtime costs more per hour than the subscription.
- Multi-region checking — genuinely useful at scale; for one site with one server it mostly generates false alarms from a flaky probe on another continent.
- Synthetic transaction monitoring — scripted logins and checkouts are powerful and need maintaining. Skip until you have someone whose job includes maintaining them.
- Status page subscriptions — a public status page is nice for customers, but not a monthly cost for a five-page brochure site. Ours is included free.
Write the two-minute runbook
The most valuable thing you can prepare is a short document, saved somewhere you can reach on a phone, listing: who hosts the site and the support number; who controls DNS and the login; who controls the domain registration; where the backups are and when they last ran; and who to call. During an outage, half the delay is invariably spent working out who to ask.
When the alert arrives
- Confirm it independently — run the status check; it names the failing stage.
- Check whether it is DNS, the server, or the certificate. Those go to three different suppliers.
- Tell customers before they tell you. A short note on social media buys enormous patience.
- Once it is fixed, write down what happened. The second occurrence of the same outage is the expensive one.