Uptime percentage is (total time minus downtime) divided by total time, multiplied by 100. The only thing that makes it fiddly is agreeing on the two inputs. Total time is the reporting period, and the usual convention is a 30-day month and a 365-day year. Downtime is where the arguments happen: whether partial degradation counts, whether scheduled maintenance is excluded, and who decides when the clock started.
Worth knowing
A worked example. A 30-day month is 43,200 minutes. If you were down for 45 minutes, uptime is (43,200 minus 45) divided by 43,200, which is 0.998958, or 99.8958%. That is a 99.9% SLA missed, by about two minutes.