Uptime / SLA Downtime Calculator (nines)

Understanding the uptime and SLA downtime calculator — concept and principle

Service level agreements (SLAs) and internal reliability targets are usually written as a percentage of availability, such as 99.9% ("three nines"). Percentages that look almost identical hide very different amounts of allowed downtime. This calculator converts an availability figure into the maximum outage time per year, month, week and day.

How it is calculated

Allowed downtime = period length × (1 − availability ÷ 100).

• Year: 365.25 days × 24 × 60 = 525,960 minutes (365.25 averages in leap years).
• Month: 30.44 days, the average month length (365.25 ÷ 12).
• Week: 7 days; day: 1,440 minutes.

Results of 60 minutes or more are shown as hours and minutes, smaller ones in minutes with two decimals.

The principle

Availability is the fraction of time a service is usable, so its complement, 1 − A, is the fraction of time it may be down. Multiplying that fraction by any period gives the downtime budget for that period. Each extra nine divides the budget by ten: 99% allows about 3.65 days a year, 99.9% about 8.77 hours, 99.99% about 52.6 minutes and 99.999% about 5.26 minutes. Engineers call this the error budget: time that can be spent on deployments, maintenance and incidents before the target is missed.

Worked example

1. Input: availability 99.9%, so 1 − 0.999 = 0.001.
2. Year: 525,960 × 0.001 = 525.96 min = 8 h 45.96 min, shown as 8 h 46 min.
3. Month: 30.44 × 1,440 × 0.001 = 43.83 min.
4. Week: 10,080 × 0.001 = 10.08 min.
5. Day: 1,440 × 0.001 = 1.44 min.

How to read the result

Ask whether your team could detect, diagnose and fix a typical incident within the monthly budget. With 43 minutes a month, one slow recovery can use it up. When services depend on each other in series, multiply their availabilities: two components at 99.9% give about 99.8% together. Redundant components in parallel do better: two independent 99.9% copies give 1 − 0.001² = 99.9999%, assuming failures are truly independent, which they often are not.

Common mistakes

• Assuming the SLA counts every outage; many contracts exclude planned maintenance or measure per month only.
• Promising more nines than the providers you depend on offer.
• Ignoring partial outages, such as slow responses or errors for some users.

Frequently asked questions

How much downtime is 99.99% per year?

About 52.6 minutes a year, or roughly 4.4 minutes per average month.

Is monthly or yearly measurement better?

Monthly windows catch problems sooner, but a single long outage can break the monthly target even if the year looks fine.

What is the difference between SLA, SLO and SLI?

An SLI is the measured indicator, an SLO is the internal target, and an SLA is the external contract, often with penalties, usually set looser than the SLO.

Sources: B. Beyer, C. Jones, J. Petoff and N. R. Murphy (eds.), Site Reliability Engineering, O'Reilly Media, 2016