Reverse Calculator
Enter your allowed downtime budget per year to find the required availability %.
How it works
An availability figure like "three nines" is only meaningful once you convert it into time a service is permitted to be down. This tool does that conversion in both directions.
Availability % is the single input: the percentage of a period during which a service is expected to be operational, as stated in a Service Level Agreement (SLA). The allowed-downtime factor is simply the complement of that number, the fraction of time the SLA permits outage. Multiply that factor by the length of any period (in seconds) and you get the downtime budget for that period.
allowed_downtime = downtime_factor × period_seconds
nines = −log₁₀( downtime_factor )
The "nines" badge reports the leading nines in the availability figure using a base-10 logarithm, so it shows fractional nines too (99.95% is 3.3 nines, not a clean count). The reverse calculator inverts the same relationship: give it an annual downtime budget in minutes and it returns the availability you would have to guarantee, using a fixed year of 31,556,952 seconds (365.24 days).
The period lengths are fixed constants: day = 86,400 s; week = 604,800 s; month = 2,629,746 s (30.44 d); quarter = 7,889,238 s; year = 31,556,952 s.
Worked example. At 99.9% availability, the factor is 1 − 0.999 = 0.001. Applied to a year: 0.001 × 31,556,952 = 31,557 seconds, which formats to 8h 45m 57s of permitted downtime per year, or about 43m 50s per month. The badge reads 3.0 nines because −log₁₀(0.001) = 3.
Benchmarks & reference points
The mapping from nines to downtime is exact arithmetic, not an estimate. The table below is derived directly from the formula above using this tool's own period constants, so you can sanity-check any SLA against it.
| Availability | Nines | Downtime / year | Downtime / month |
|---|---|---|---|
| 99% | 2.0 | 3d 15h 39m 30s | 7h 18m 17s |
| 99.5% | 2.3 | 1d 19h 49m 45s | 3h 39m 9s |
| 99.9% | 3.0 | 8h 45m 57s | 43m 50s |
| 99.95% | 3.3 | 4h 22m 58s | 21m 55s |
| 99.99% | 4.0 | 52m 36s | 4m 23s |
| 99.999% | 5.0 | 5m 16s | 26s |
Related planning conventions are worth pairing with an SLA rather than reading off it: recovery-time and recovery-point objectives (RTO/RPO) are typically tiered, e.g. minutes for tier-0 systems, hours for tier-1, a day or more for lower tiers. Underlying availability itself is often modelled from reliability data as MTBF ÷ (MTBF + MTTR), where MTBF is mean time between failures and MTTR is mean time to repair. These are conventions, not universal standards; confirm the tiers your own organisation uses.
Using this in the boardroom
The value of this conversion is that it turns an abstract percentage into a number a board can weigh against revenue and risk. "99.9%" invites a nod; "up to 8 hours and 46 minutes of outage a year, and we have no committed remedy under four hours" invites a decision. When you present an availability target, show the downtime budget beside it and state whether current architecture and staffing can actually hold that budget, the gap between the contractual number and the operational reality is usually the real story.
Pair the downtime budget with three companions: the RTO/RPO the business has agreed for the same system, the historical outage minutes you have actually incurred, and the cost of the marginal nine. Each additional nine is roughly a tenfold reduction in permitted downtime and typically a step-change in cost (redundancy, automation, on-call depth). Framing the ask as "what does the next nine buy, and what does it cost" keeps the conversation grounded in expected loss rather than aspiration.
The one takeaway for a board slide
An SLA percentage is a downtime budget in disguise. State the target, the minutes it permits, the minutes you incurred last year, and whether you have the redundancy to close the gap, that is a resilience business case in four numbers.
Frequently asked questions
What does "five nines" actually mean in downtime?
Five nines is 99.999% availability, which permits only about 5 minutes and 16 seconds of downtime per year, roughly 26 seconds per month. Each added nine cuts the permitted downtime by about a factor of ten, so the jump from four nines to five nines is a tenfold tightening of your outage budget.
Does an SLA percentage include planned maintenance?
It depends entirely on the contract wording. Many vendor SLAs exclude scheduled maintenance windows, force majeure, and customer-caused outages from the calculation, which can make a headline "99.9%" far more generous than it appears. Always confirm what counts as downtime and over what measurement window before comparing SLAs.
Is availability measured monthly or annually?
Both are common, and the window changes the practical meaning. A monthly SLA resets the budget every month, so a bad month cannot be averaged away; an annual figure gives more slack but hides clustered outages. This tool shows every period side by side so you can see the budget under whichever window your contract uses.
How is availability different from RTO and RPO?
Availability is how much total downtime is permitted over a period; RTO (recovery time objective) is how fast a specific outage must be resolved; RPO (recovery point objective) is how much data loss is acceptable. A service can meet an annual availability target while still breaching an RTO on a single long incident, which is why the three are set and reviewed together.
Why doesn't the tool show a clean number of nines for 99.95%?
The badge uses a base-10 logarithm rather than counting literal "9" characters, so it reports fractional nines: 99.95% comes out as 3.3 nines. This is more honest than rounding, because 99.95% genuinely sits between three and four nines in downtime terms.