Uptime / SLA Downtime Calculator

About this tool

Pick an uptime target — 99%, 99.9%, 99.95%, 99.99%, 99.999% or any custom percentage — and the calculator converts it into the downtime that target permits per day, week, month, quarter and year. The arithmetic is simple: the unavailable fraction is one minus the target, applied to the seconds in each period. Everything runs in your browser.

Its value is in making an abstract percentage concrete. Three nines sounds close to perfect until you see it permits about 8 hours 46 minutes a year, which one bad Saturday morning consumes entirely. Four nines allows roughly 52 minutes 36 seconds across the whole year — less than a single unplanned reboot with a slow filesystem check. Seeing those numbers before signing a contract or promising a target tends to change the conversation about redundancy and failover testing. If the target you need is beyond what your current design supports, that is a disaster recovery problem rather than a monitoring one.

How to use it

Select a target, or choose Custom and type your own percentage, then read the table:

• Compare the per-year row against your actual downtime over the last twelve months before committing.
• Use the per-month row if your contract is measured monthly, as most provider SLAs are.
• Check the allowance covers planned maintenance, unless the contract excludes it.
• Remember dependencies compound: a service built on three components each at 99.9% cannot itself reach 99.9% without redundancy.

Common questions

How much downtime does 99.9% actually allow?

About 8 hours 46 minutes a year, or roughly 43 minutes 50 seconds a month. Four nines (99.99%) allows about 52 minutes 36 seconds a year, and five nines (99.999%) about 5 minutes 15 seconds. Each additional nine cuts the allowance by a factor of ten, which is why the jump from three nines to four is an architecture change rather than a tuning exercise.

Does planned maintenance count against the SLA?

That depends entirely on the contract, not on the maths. Most provider SLAs exclude announced maintenance windows, force majeure and customer-caused faults, and many measure per calendar month rather than per year. Read the measurement window and the exclusion list before comparing two agreements on their headline percentage.

Why is the monthly figure not exactly one twelfth of the yearly one?

Calendar months differ in length, so this tool uses an average month of 30.4375 days — a 365.25-day year divided by twelve. The per-month row is therefore an average rather than a specific month. A contract measured over February allows less absolute downtime than the same percentage measured over July, and yearly figures here sit a few seconds above the commonly quoted 365-day values.