Enter an uptime target and see exactly how much downtime it allows. Or go the other way: enter the downtime you had and get the uptime you actually delivered.
Runs entirely in your browser — nothing you enter here is sent to a server.
An SLA figure looks abstract until you convert it to time. The arithmetic is simple — allowed downtime is the length of the period multiplied by (100 − SLA) / 100 — but the results are easy to misjudge. The step from 99.9% to 99.99% sounds small and removes nine tenths of your margin: about 43 minutes a month becomes about 4 minutes.
This calculator uses an average month of 30.44 days and a year of 365.25 days, which is what most providers mean when the contract does not say otherwise. If your agreement defines a billing month or excludes maintenance windows, the period you measure against is shorter and so is the allowance.
Roughly 7 hours 18 minutes a month. Fine for internal tools and batch systems; noticeable for anything customers use daily.
About 43 minutes a month. The usual target for SaaS products and what most cloud services promise per instance.
About 4 minutes a month. A single slow deploy can spend it, so it needs redundancy and automated failover.
About 26 seconds a month. No human can respond in that time; recovery has to be fully automatic.
Choose the number from the downtime, not the other way round. Ask how long a real incident takes you from first alert to recovery, how many of those you have in a typical quarter, and whether planned maintenance counts. If one ordinary incident costs you 30 minutes, a 99.99% promise is already broken the first time it happens.
Remember that an SLA is only as good as its measurement. Uptime measured from inside your own network misses DNS, certificate and routing failures that your customers do see, which is why availability is normally measured by external checks at a fixed interval.
About 1 minute 26 seconds a day, 10 minutes a week, 43 minutes 50 seconds a month and 8 hours 46 minutes a year.
An SLO is the internal target your team aims for. An SLA is the contractual promise to customers, usually with credits or penalties attached. Teams normally set the SLO stricter than the SLA so there is a safety margin.
Only if your agreement says so. Many SLAs exclude announced maintenance windows, which makes the published percentage look better than the availability users experience. Check the definition before comparing providers.
30.44 days, the average over a year. A 31-day month allows slightly more downtime for the same percentage and February slightly less.
You need measurements from outside your infrastructure at a regular interval. An uptime monitor records every failed check, so the downtime figure for the period is a fact rather than an estimate.
Upzilla measures real uptime from outside your network and gives you the report to prove it. 10 monitors free, no card required.