What is uptime monitoring? The basics, explained simply.

How uptime monitoring works

A monitoring service requests your website at fixed intervals from the outside, just like a visitor would. It measures whether the site responds, how fast it responds and whether the response is actually correct. If the answer fails or returns an error, you get alerted, via email, Slack, Discord or webhook.

The key phrase is "from the outside": the checks run on separate infrastructure, not on your own server. Monitoring that runs on the same machine as the site it watches goes down together with it, exactly when you need it.

The main check types

In practice, four check types cover almost every case:

• HTTP(S) check: requests a URL and validates the status code (e.g. 200 OK) and response time. The standard for websites, shops and APIs.
• Keyword check: additionally verifies that an expected text appears in the response. Catches "soft failures" where the server returns 200 but the page shows an error message or nothing at all.
• SSL certificate check: warns before a certificate expires. Expired certificates are one of the most common and most avoidable causes of downtime.
• Heartbeat / cron job check: reverses the direction. Your job reports in after each successful run; if the report stays out, you get alerted. Ideal for backups, imports and any recurring background job.

What matters when choosing a service

Services differ less in the "if" than in the "how". These points matter day to day:

• Check interval: every 5 minutes is fine for most sites; 60 seconds makes sense for shops and business-critical services.
• False-alarm prevention: a good service re-checks before alerting. One lost request at 3 a.m. should not wake you up.
• Alert channels: email is table stakes; Slack/Discord/webhooks help teams. Recovery alerts matter too.
• Evidence: if you need to prove to clients that checks ran (or that the site was up), you need exportable, ideally signed logs instead of screenshots.
• Provider and data location: monitoring involves URLs, error details and response times of your systems. Where that data lives and who your contract partner is, is a privacy question. More in our GDPR article.

Hosted service or self-hosted?

Open-source tools like Uptime Kuma are mature and a good fit for homelabs. For business-critical sites, self-hosting has a structural catch: you have to run, update, back up and monitor the monitoring yourself, and if it shares your infrastructure, it shares your outages.

The pragmatic answer for many teams: run your own tool internally if you like, and use an external, independent service for the sites your customers see.

What does uptime monitoring cost?

The range goes from permanently free tiers (usually a few monitors at 5-minute intervals) to 5 to 30 euros per month for small teams, up to enterprise platforms costing hundreds per month.

For context: Clesk Uptime includes 3 monitors free forever; Pro is €79 per year including VAT (20 monitors, 60-second checks). In the end, compare it to what a single unnoticed outage would cost.

Frequently asked questions