Uptime is the amount of time an IT system, server, network device, application, or other service operates continuously without failures or unplanned interruptions. The term is also used to describe a system’s availability level over a specific period, expressed as a percentage.
For example, 99.9% uptime means that a service should remain available for 99.9% of the measured time. The higher this percentage, the less downtime is allowed.
What Uptime Indicates
Uptime reflects the stability and availability of IT infrastructure. The metric is used for servers, websites, cloud platforms, storage systems, network equipment, and other components.
Uptime may refer to:
- continuous operating time — for example, a server has been running without a reboot for 180 days
- availability percentage — for example, service availability for a month was 99.95%
- an SLA target — a provider commits to a certain availability level, such as 99.99%
Therefore, context is important when using the term. Uptime as a duration of operation and uptime as an availability percentage are related but not entirely identical metrics.
How Uptime Is Calculated
When uptime is used as an availability metric, it is calculated as the ratio of normal system operating time to the total measured time:
Uptime = (Total Time − Downtime) / Total Time × 100%
For example, a system is expected to operate for 720 hours over a 30-day period. If the total counted downtime is 3 hours:
Uptime = (720 − 3) / 720 × 100% ≈ 99.58%
In an SLA, Agreed Service Time (AST) may be used instead of the entire calendar period for the calculation. In addition, certain types of downtime, such as scheduled maintenance, may be excluded from the calculation. The specific rules are defined by the terms of the agreement.
What 99%, 99.9%, and 99.99% Uptime Mean
Small differences in uptime percentages have a significant impact on the amount of allowable downtime. For a 30-day period (720 hours), the percentages approximately correspond to the following values:
- 99% — up to 7 hours 12 minutes of downtime
- 99.9% — up to 43 minutes 12 seconds
- 99.99% — up to 4 minutes 19 seconds
- 99.999% — about 26 seconds
The last figure is commonly referred to as “five nines.” The higher the required uptime, the more fault-tolerant the architecture generally needs to be to achieve it.
Uptime and Downtime
Downtime is a period during which a system or service is unavailable or unable to perform its intended functions.
Uptime and downtime are used together when evaluating availability:
Total Measured Time = Uptime + Downtime
However, not every technical incident is necessarily counted as downtime. An SLA typically defines the criteria for unavailability, measurement methods, and exclusions from the calculation.
For example, scheduled maintenance may not be counted as downtime if it takes place within a pre-agreed maintenance window.
Uptime and SLA
An SLA (Service Level Agreement) defines a provider’s service level commitments. Uptime or availability may be one of the metrics specified in such an agreement.
For example, an SLA may specify 99.99% monthly availability. If the actual figure falls below the agreed level, service credits, compensation, or other remedies specified in the agreement may apply.
When comparing SLAs from different providers, looking only at the number of “nines” is not enough. It is also important to consider:
- the period over which the metric is calculated
- what time period is included in the calculation
- what qualifies as unavailability
- whether scheduled maintenance is excluded
- which events are treated as exceptions
- how the start and end of downtime are recorded
Therefore, two offers with the same stated uptime of 99.9% may have different availability terms in practice.