How to use it
- Enter the total operating time in hours: the time the system was running during the period you observe.
- Enter the number of failures in that period and the total repair time in hours.
- Optionally change the mission time to see the chance of running that long without a failure.
Formulas
- MTBF = operating time ÷ failures
- MTTR = repair time ÷ failures
- Availability A = MTBF ÷ (MTBF + MTTR)
- Failure rate λ = 1 ÷ MTBF, also shown per 10⁶ hours
- Reliability R(t) = e^(−t ÷ MTBF), assuming a constant failure rate
- Downtime per year = (1 − A) × 8,760 h
Worked example
A server ran 8,760 h in a year, failed 4 times and was down for 12 h in total.
- MTBF = 8,760 ÷ 4 = 2,190 h; MTTR = 12 ÷ 4 = 3 h.
- Availability = 2,190 ÷ 2,193 = 99.8632 %, or 11.98 h of downtime a year.
- λ = 1 ÷ 2,190 = 0.00045662 /h, which is 456.62 failures per million hours.
- The chance of a 30-day (720 h) month without failure: e^(−720 ÷ 2,190) = 71.98 %.
Tips
- Use enough history. Four failures give a rough estimate; the more failures in the record, the more reliable the MTBF.
- Keep units consistent. Enter everything in hours; for days multiply by 24 first.
- Availability targets such as 99.9 % allow 8.76 hours of downtime a year; the uptime SLA calculator turns any percentage into allowed downtime.