Business7 min read

Uptime SLAs Explained for UK SMEs: What to Promise and How to Prove It

Service Level Agreements on uptime require you to measure what you are promising. This guide explains 99.9% vs 99.99% uptime, how SLAs are calculated, and how to back them with real monitoring data.

Uptime SLAs sound straightforward — 99.9% availability — but the numbers hide significant complexity. Promising 99.9% uptime to a client means committing to less than 9 hours of downtime per year. Without monitoring data to back that commitment, it is an empty promise. With the right tooling, it becomes a genuine differentiator.

What the numbers actually mean

99% uptime allows 87.6 hours of downtime per year — more than 3.5 days. 99.9% ("three nines") allows 8.76 hours per year. 99.99% ("four nines") allows just 52 minutes. 99.999% ("five nines") allows 5 minutes. Most SME web services operate at somewhere between 99% and 99.9% without realising it. The gap between three nines and four nines is the difference between a planned maintenance window and an enterprise-grade infrastructure investment.

What counts as downtime in an SLA

This is where SLA negotiations get granular. Planned maintenance windows are typically excluded — you notify clients in advance and the window does not count against your uptime percentage. Partial outages (degraded performance, high error rates, single region failures) may or may not count depending on how the SLA is drafted. The cleaner your monitoring data, the cleaner your SLA disputes become.

Using monitoring data to evidence SLA compliance

If a client claims your service was down on a specific date and time, your monitoring logs are your evidence. A tool that records the outcome of every check, with timestamps, gives you an audit trail. VP Watchtower stores check results for the 7-day showcase window — for permanent evidence you need a long-term monitoring solution.

What to include in an uptime SLA

A well-drafted uptime SLA for a UK SME should specify: the measurement window (monthly vs annual), what constitutes a service failure, planned maintenance exclusions, the notification procedure for incidents, remedies for SLA breach (service credits, compensation caps), and how uptime is calculated and reported. Get your solicitor to review anything that carries financial remedies.

Starting with a realistic baseline

Before committing to an SLA figure, run monitoring for 30 days to establish your actual baseline. Use that data to set a figure you can genuinely exceed, not one that will cause you to breach on the first bad week. VP Watchtower gives you a 7-day window to start building that picture for a specific service.

Try it free — 7 days, no billing

Create a hosted status page with HTTP, TCP, DNS, SSL, and SEO monitoring. Live in under 5 minutes.

Create your status page →