"99.9% uptime" sounds precise, but the number alone tells you almost nothing without knowing how it's measured, over what period, and what happens when it's missed. Vendors can present very different realities under the same headline percentage.
How uptime is actually measured
Uptime SLAs are typically calculated over a defined period — monthly is most common — as the percentage of time a system was available against total possible time. The critical variable is what counts as "down": some definitions only count a full outage, while others count degraded performance or partial service loss. A 99.9% SLA that only counts complete outages can mask a system that's frequently slow but never technically "down."
Typical credits and penalties
Most SLAs specify a service credit — a percentage discount on the next invoice — when uptime falls below the committed threshold, scaled to how far below. It's worth checking whether credits are the sole remedy (meaning there's no path to terminate for repeated failures) or whether persistent breaches also trigger termination rights.
Questions to ask before trusting a number
- What specifically counts as downtime — full outage only, or degraded performance too?
- Over what measurement window is the percentage calculated, and is that window disclosed proactively or only on request?
- Are scheduled maintenance windows excluded from the calculation, and if so, how much maintenance time is typical?
- Is uptime reporting available to you directly, or do you have to rely solely on the vendor's self-reported figure?
- What's the actual remedy — credit only, or does sustained failure open a path to exit the contract?