See what 99.9% or 99.99% really allows in minutes, work backwards from downtime to an uptime figure, and combine components in series or parallel.
Runs entirely in your browser. Nothing you enter is sent to a server.
Uptime to downtime
Allowed downtime for the chosen uptime
Period
Allowed downtime
Downtime to uptime
Combined availability of components
Assumes components fail independently. Shared dependencies such as a common region or database reduce the real figure.
How it works
Availability is the share of a period during which a service works. Allowed downtime is therefore period × (1 − uptime ÷ 100). This tool uses 24-hour days, a 7-day week, a 30-day month and a 365-day year, which are the usual conventions for quick comparisons. Your contract may define the measurement window differently.
For a request that must pass through several components, availability multiplies: A = A₁ × A₂ × …. For redundant components where any one can serve the request, unavailability multiplies instead: A = 1 − (1 − A₁)(1 − A₂)…. Both formulas assume the components fail independently.
Common uses
Checking whether an SLA promise is realistic for the architecture behind it.
Explaining to stakeholders what “three nines” and “four nines” mean in minutes.
Estimating how much a load balancer, extra region or standby database improves availability.
Setting an error budget for an SLO.
Frequently asked questions
How much downtime does 99.9% uptime allow?+
Over a 365-day year, 99.9% allows about 8 hours 45 minutes of downtime, roughly 43 minutes in a 30-day month and about 1 minute 26 seconds per day. The calculator shows the exact figures for any percentage.
Why is a chain of components less available than each one?+
Components in series all have to work for the service to work, so their availabilities multiply. Two components that are each 99.9% available give 99.9% × 99.9% = 99.8001% together. Every extra dependency in the request path lowers the total.
How does redundancy improve availability?+
Components in parallel only fail when all of them fail at the same time, so the combined availability is 1 − (1 − A₁)(1 − A₂)… Two independent 99% components in parallel give 99.99%. This assumes failures are independent, which shared causes such as a common region, deployment or database can break.
Does an SLA percentage count planned maintenance?+
That depends entirely on the contract. Many SLAs exclude scheduled maintenance, and each has its own measurement window, exclusions and remedies. Read the definitions in the agreement; this tool only does the arithmetic.
What is the difference between an SLA, an SLO and an SLI?+
An SLI is a measured indicator, such as the share of successful requests. An SLO is the internal target for that indicator. An SLA is the contractual promise to a customer, usually with remedies if it is missed and usually set looser than the internal SLO.
These tools are free to use. If you need cloud and DevOps, web or mobile development, CRM integrations or managed IT support, tell us what you are working on.
