Server Uptime SLAs Explained: What 99.9% and 99.99% Really Mean

Server Uptime SLA

Hosting providers, data centres and IT service companies all promise uptime in their service level agreements (SLAs): 99.9%, 99.99% and so on. Those numbers sound similar but mean very different amounts of downtime, and the small print often matters more than the headline figure. This guide explains what uptime guarantees really mean and how to evaluate them.

Quick answer: an uptime percentage tells you the maximum downtime the provider is allowed before it owes you anything. 99.9% allows about 8.8 hours of downtime a year; 99.99% allows about 53 minutes. Check how uptime is measured, what is excluded (especially planned maintenance), and what compensation you actually get, which is usually a service credit rather than payment for your losses.

What uptime percentages mean

UptimeDowntime per yearDowntime per 30-day month
99%About 3.65 daysAbout 7.2 hours
99.5%About 1.83 daysAbout 3.6 hours
99.9%About 8.8 hoursAbout 43 minutes
99.95%About 4.4 hoursAbout 22 minutes
99.99%About 53 minutesAbout 4.3 minutes
99.999%About 5.3 minutesAbout 26 seconds
Good to know: each extra “9” cuts allowed downtime by a factor of ten, and the cost of delivering it rises steeply. Match the SLA to what downtime actually costs your business rather than buying the highest number available.
SLA commitment and contract terms

What to check in an SLA

  • Measurement: is uptime measured monthly or yearly, and by whom? Monthly measurement is stricter.
  • Exclusions: planned maintenance, customer-caused issues, third-party networks and force majeure are often excluded.
  • Scope: does the SLA cover the network, the power, the server hardware, or the whole service?
  • Response and fix times: how fast will someone respond to and resolve a hardware fault?
  • Remedies: what credit do you get, how do you claim it, and is there a cap?
Watch out: most SLA remedies are service credits, often capped at a share of your monthly fee. They rarely cover your lost revenue. Your own redundancy and backups are what actually protect the business.

Building your own uptime

An SLA from a provider covers their part of the chain. For servers you run yourself, uptime comes from design: redundant power supplies and RAID inside each server, two or more servers for critical services, UPS protection for load shedding, and monitoring. Read our guides to high availability architectures and disaster recovery planning.

Pro tip: when buying hardware, check the supplier warranty and how quickly replacement parts can be supplied. Keeping a spare drive and power supply on site often shortens real-world downtime more than any SLA clause.

Our own terms for refurbished hardware are explained in our refurbished server warranty guide.

Frequently asked questions

What does 99.9% uptime mean?

It means the service may be unavailable for up to 0.1% of the time, roughly 8.8 hours a year or about 43 minutes in a 30-day month, before the SLA is breached.

Is 99.99% uptime realistic?

Yes, for well-designed services with redundancy at every layer, but it costs significantly more than 99.9%.

What happens if a provider breaks its SLA?

Usually you can claim a service credit against future bills, subject to the SLA’s conditions and caps.

Does planned maintenance count as downtime?

Often not. Many SLAs exclude scheduled maintenance, so check the definition carefully.

Building redundancy into your own setup? Browse our refurbished servers or talk to our team.

Server Site