Bookkeeping for outages: MTBF (mean time between failures) is the average uptime between incidents, MTTR is the average recovery time, and inherent availability is MTBF / (MTBF + MTTR). Enter incidents to get all four, or start from an availability SLO and see how much downtime it buys you across common windows. Everything runs locally.
| Nines | Availability | Downtime / year |
|---|---|---|
| 1 | 90 % | 36.5 d |
| 2 | 99 % | 3.65 d |
| 2.5 | 99.5 % | 1.83 d |
| 3 | 99.9 % | 8h 45m |
| 3.5 | 99.95 % | 4h 22m |
| 4 | 99.99 % | 52m 34s |
| 4.5 | 99.995 % | 26m 17s |
| 5 | 99.999 % | 5m 15s |
| 6 | 99.9999 % | 31.5s |
MTBF — Mean time between failures. Only applies to repairable systems. Two conventions exist: uptime-only, or the full cycle including repair.
MTTF — Mean time to failure, for non-repairable components (a disk that dies). MTBF ≈ MTTF + MTTR.
MTTR — Mean time to repair. Usually measured wall-clock from detection to full recovery. Some teams distinguish MTTD (detect), MTTA (acknowledge), MTTR (respond / restore) and MTTF (fix root cause).
Inherent availability — A = MTBF / (MTBF + MTTR). Ignores planned maintenance and logistics delay; use operational availability if you need to include those.