#mtbf

Reliability-engineering maths as an API, computed locally and deterministically — the availability, MTBF and failure maths behind SLAs and dependable systems. The availability endpoint converts between MTBF and MTTR, a target availability and the SLA "nines": give it a mean time between failures and a mean time to repair and it returns the availability A = MTBF/(MTBF+MTTR) and the downtime per year, month, week and day; give it a number of nines and it returns the budget — three nines (99.9 %) is 8.76 hours of downtime a year, five nines (99.999 %) just 5.26 minutes. The reliability endpoint computes the probability a unit survives a mission time under the exponential model R(t) = e^(−λt) with its constant hazard λ = 1/MTBF, or the Weibull model R(t) = e^(−(t/η)^β) — β below one for infant mortality, one for random failures, above one for wear-out — returning the reliability, failure probability, hazard rate and the mean life η·Γ(1+1/β). The system endpoint combines component reliabilities into a system: series (the weakest link, ΠRᵢ), parallel redundancy (1−Π(1−Rᵢ)) or k-of-n voting. Everything is computed locally and deterministically, so it is instant and private. Ideal for SRE, DevOps, hardware-reliability, safety-engineering and SLA-planning app developers, uptime-budget and redundancy-design tools, and engineering education. Pure local computation — no key, no third-party service, instant. Live, nothing stored. 3 compute endpoints. This is reliability and availability maths; for queue wait-times use a queueing API and for live uptime checks use a monitoring service.

api.oanor.com/reliability-api