Can you count on a server, a pump or a production line? Two numbers answer that: the MTBF, the average time between failures, and the MTTR, the average time to repair. Together they give the availability, the share of time that the system works. This calculator computes the availability, the downtime per year, the expected failures, the number of “nines” and the reliability over a mission.
How the calculation works
- Availability = MTBF ÷ (MTBF + MTTR)
- Downtime per year = (1 − availability) × 8,760 hours
- Failures per year = 8,760 ÷ (MTBF + MTTR)
- Number of nines = −log₁₀(1 − availability)
- Reliability over a mission of t hours = e^(−t ÷ MTBF)
Example
A machine has an MTBF of 2,000 hours and an MTTR of 4 hours. Its availability is 99.80%, about 2.7 nines. It is down for 17.5 hours a year, with 4.4 failures a year. Over a one-week mission (168 hours), the probability of no failure is 91.9%.
Why it matters
Availability targets drive maintenance budgets, spare parts and service contracts. The formula shows two ways to improve: fail less often, or repair faster. Cutting the MTTR in half has the same effect as doubling the MTBF when repairs are short, and is often much cheaper.
Good to know
The formulas assume a constant failure rate and a repair time that is independent of the failure. Real equipment has early failures, wear-out and logistics delays, and planned maintenance also takes machines out of service. For service-level agreements with a monthly target, see the uptime SLA calculator. A system made of several parts in series is available only when all of them are, so its availability is the product of the availabilities of its parts; redundant parts in parallel raise it. A few repeated measurements of repair times are worth more than a single estimate. This tool is for education.
Related tools
The uptime SLA calculator converts a percentage into allowed downtime, the Little’s law calculator relates work in progress to delivery time, and the server capacity calculator sizes a service.