redzilla
All tools
Datacenter

Availability & SLA

Model availability from MTBF/MTTR or per-component percentage and its redundancy (series, parallel or N+1). Get the nines and the downtime per year, month and week.

Topology

Series: all must work; availability drops. A = ∏ Aᵢ

Per-component input

A = MTBF / (MTBF + MTTR), in the same time units (hours).

Components
redzilla.cl — sla
 
Availability
Downtime / year
out-of-service time
Downtime / month
monthly average
Downtime / week
weekly average
A (%)
Nines
Year
Per-component contribution
Component A (%) Nines Downtime / year

The nines of availability
SLA Nines Downtime / year Downtime / month Downtime / week

Periods: year = 525,600 min (365 d), month = 43,800 min (year/12), week = 10,080 min (7 d). MTBF/MTTR must be in the same time units. The model assumes independent, uncorrelated failures.

Runs locally in your browser · no sign-up · nothing leaves your browser

How it works

The calculator models the availability of a system from its components. Each component is entered either via MTBF/MTTR (A = MTBF / (MTBF + MTTR), in the same time units) or directly as a percentage. It then combines them by topology: in series everything must work (A = ∏ Aᵢ, availability drops), in parallel one unit is enough (A = 1 − ∏(1 − Aᵢ), availability rises) and for k-of-n with identical units it uses the binomial sum A = Σ C(n,i)·aⁱ·(1−a)ⁿ⁻ⁱ for i from k to n, which covers N+1 schemes.

With the availability solved it reports the nines (−log₁₀(1 − A)) and the expected downtime per year (525,600 min), month (43,800 min) and week (10,080 min): downtime = (1 − A) × period. The model assumes independent, uncorrelated failures, the standard assumption in reliability math; a common-cause failure (power, fire) can take down redundant branches at once.

Example: server + switch + WAN link in series

  1. Per-component availability: server 50000/(50000+4) ≈ 99.9920 %, core switch 200000/200002 ≈ 99.9990 %, WAN link 20000/20006 ≈ 99.9700 %.
  2. In series they multiply: A ≈ 99.961 %, about 3.4 nines.
  3. Annual downtime: (1 − 0.99961) × 525600 ≈ 205 min, roughly 3 h 25 min per year, dominated by the WAN link.

Frequently asked questions

How many minutes of downtime per year does a 99.9 % SLA allow?
99.9 % (three nines) allows 525.6 minutes per year, about 8.76 hours. At 99.99 % that drops to 52.6 minutes a year and at 99.999 % (five nines) to 5.26 minutes. Every additional nine divides the downtime by ten.
What is the difference between components in series and in parallel?
In series the system fails if any component fails, so the unavailabilities add up: the chain is always worse than its worst link. In parallel the system fails only if all units fail: two 99 % units in parallel reach 99.99 %, because the probability of simultaneous failure is 0.01 × 0.01.
What does N+1 redundancy mean and how do I enter it?
N+1 means having one spare unit on top of the N you need: the system works while at least k = n − 1 units are running. In the calculator pick the k-of-n topology and enter n total units and k required; for example 3 UPS units where 2 are enough is a 2-of-3 (N+1).
What happens if MTBF and MTTR are in different units?
The result comes out wrong: the formula A = MTBF/(MTBF+MTTR) requires the same unit for both, normally hours. An MTBF of 50,000 h with an MTTR of 4 h gives 99.992 %; if the MTTR were entered in minutes as 240 fake hours, availability would artificially drop to 99.52 %. Everything runs in your browser and no data is sent anywhere.
Was this tool useful?
Disclaimer We take great care to keep every tool accurate and review it thoroughly; even so, we can't guarantee it is free of errors or take responsibility for how the results are used. We recommend double-checking anything critical.
Found an error? Let us know →