Free tool
SLA and uptime calculator
See how much downtime an SLA allows per day, week, month and year, what availability you actually hit after an outage, and what your dependencies do to your own SLA. Free, no sign-up, nothing leaves your browser.
Uptime and downtime by number of nines
Allowed downtime for common availability targets. A month is 30.44 days and a year 365.25 days.
| Availability | Daily | Weekly | Monthly | Yearly |
|---|---|---|---|---|
| 90% one nine | 2h 24m | 16h 48m | 3d 1h 3m | 36d 12h 36m |
| 95% one and a half nines | 1h 12m | 8h 24m | 1d 12h 31m | 18d 6h 18m |
| 99% two nines | 14m 24s | 1h 40m 48s | 7h 18m 18s | 3d 15h 39m |
| 99.5% two and a half nines | 7m 12s | 50m 24s | 3h 39m 9s | 1d 19h 49m |
| 99.9% three nines | 1m 26s | 10m 5s | 43m 50s | 8h 45m 58s |
| 99.95% three and a half nines | 43s | 5m 2s | 21m 55s | 4h 22m 59s |
| 99.99% four nines | 8.6s | 1m | 4m 23s | 52m 36s |
| 99.999% five nines | 0.86s | 6.0s | 26s | 5m 16s |
| 99.9999% six nines | 0.09s | 0.60s | 2.6s | 32s |
SLA, SLO, SLI and error budgets
SLA
Service level agreement: the availability you promise customers, usually in a contract with service credits if you miss it. Example: 99.9% monthly uptime.
SLO
Service level objective: the internal target you run the service to. It should be stricter than the SLA, for example 99.95%, so you notice trouble before customers are owed credits.
SLI
Service level indicator: the measurement itself, such as the share of successful requests or of minutes a health check passed. The SLO is a target for the SLI.
Error budget
The downtime an SLO allows, seen as a budget. At 99.9% a month you have about 43 minutes to spend on deploys, incidents and maintenance before you breach.
Where SLAs are actually won or lost
Detection. Downtime starts counting when the service breaks, not when someone notices. If the first signal is a customer email 20 minutes in, a 99.95% month (21 minutes of budget) is already gone. Monitor your own endpoints and the third-party services you depend on.
Response. Time to acknowledge is pure budget. A page that rings for five minutes before anyone answers costs more than many fixes. A clear rotation with a backup and automatic escalation keeps that number small.
Dependencies. Your SLA can never beat the product of the SLAs you rely on. Use the composite mode above to check your promise is achievable before you sign it.
Communication. Many SLAs require you to report incidents. A status page updated during the incident is also what keeps support tickets down.
Frequently asked questions
How much downtime is 99.9% uptime?
99.9% availability ("three nines") allows about 43 minutes 50 seconds of downtime a month, 10 minutes 5 seconds a week, or 8 hours 46 minutes a year.
How much downtime does 99.99% allow?
99.99% ("four nines") allows about 4 minutes 23 seconds of downtime a month and 52 minutes 36 seconds a year. At that level a single slow deploy or one paged engineer who takes ten minutes to log in can use up the whole month.
How do you calculate uptime percentage?
Uptime % = (total time minus downtime) divided by total time, times 100. For example, 45 minutes of downtime in a 30.44-day month is 1 minus 45 / 43,830 minutes, which is 99.897%: just short of 99.9%.
How do you calculate a composite SLA?
When your service needs every dependency to be up, multiply their availabilities. Three dependencies at 99.99%, 99.95% and 99.99% give 0.9999 x 0.9995 x 0.9999 = 99.93%, which is lower than any one of them. Redundancy works the other way: two independent copies at 99.9% give 99.9999%.
What is a good SLA for a SaaS product?
Most B2B SaaS products commit to 99.9% a month, and infrastructure providers to 99.95% or 99.99% for individual services. Promise less than your SLO, and only promise what your on-call process can defend: four nines means someone must be acknowledging pages within a minute or two.
Is this SLA calculator free?
Yes, free with no sign-up. Everything is calculated in your browser.
Protect the SLA you promised
Know first
Uptime checks on your own sites and endpoints, plus live status for 3,000+ services you depend on, alerting in Slack.
Answer fast
Page whoever is on call by phone, SMS and app, and escalate to the backup automatically if nobody acknowledges.
Tell customers
A public status page you can update from Slack while you fix the incident.
Put your on-call where your team already is.
Install in two minutes. Import your existing schedule. Page someone tonight.