How to use this calculator
- Enter or choose calculate from.
- The availability updates as you type. Use Copy result or Share to keep it.
Method
Availability = (Measured time − Downtime) ÷ Measured time, where Measured time = Period − Excluded maintenance; or Availability = MTBF ÷ (MTBF + MTTR); Nines = −log₁₀(1 − Availability)
- Downtime
- unplanned time the service was unusable
- MTBF
- mean time between failures
- MTTR
- mean time to recover
- Year
- 365.25 days (525,960 minutes)
Worked example
- Measured time = 30 × 1,440 = 43,200 min.
- Availability = (43,200 − 43.2) ÷ 43,200 = 99.9%.
- At that rate a year has 0.1% × 525,960 = 525.96 min (about 8.77 h) of downtime.
43.2 minutes of downtime in 30 days is 99.9% availability, or three nines.
What the result means
Compare the result with your SLA or SLO. Each extra nine cuts allowed downtime by a factor of ten, which usually needs a step change in architecture, not just faster fixes.
Assumptions and limits
- Downtime is total: partial outages count fully unless you weight them yourself.
- The MTBF/MTTR form is a steady-state average; short periods with few incidents can differ a lot from it.
Questions
How is this different from an SLA downtime calculator?
An SLA downtime calculator starts from a target such as 99.9% and tells you how much downtime it allows. This page starts from what actually happened and gives the availability achieved.
Does planned maintenance count as downtime?
It depends on your SLA. Many exclude announced maintenance; enter it in the excluded field to remove it from the measured period.