SLO Downtime Table
The "table of nines": how much downtime each availability target allows per day, week, month and year. Add your own targets, then copy it for a doc or a slide.
Runs entirely in your browser — nothing you enter here is sent to a server.
Reading the table
Each row is an availability target; each cell is the longest total outage that still meets it over that period. The figures use an average month (30.44 days) and year (365.25 days). Read across to see how an allowance that looks generous per year becomes tight per day, and read down to see what each extra nine costs: every step divides the allowance by ten.
The daily column matters more than it seems. A target is normally judged over a month or a quarter, but a single incident happens on one day — so compare the monthly allowance with how long your incidents actually last.
90 – 99%
Hours of downtime a month. Suitable for internal tools, staging environments and batch jobs that can catch up.
99.5 – 99.9%
From about 3.6 hours down to 43 minutes a month. Where most business software sits; achievable with one region and good practices.
99.95 – 99.99%
22 down to 4 minutes a month. Requires redundancy, health checks and automated failover.
99.999%
Five minutes a year. Multi-region, no manual steps in recovery, and very careful change management.
How teams use it
Put the table next to your incident history. If last quarter's incidents add up to 70 minutes, you delivered roughly 99.95% — whatever the status page says. The table is also a quick sanity check on a vendor's SLA: translate their percentage into minutes and ask whether an outage that long would be acceptable to you.
Frequently asked questions
How much downtime is "five nines"?
99.999% allows about 0.86 seconds a day, 26 seconds a month and 5 minutes 16 seconds a year.
Why does each extra nine matter so much?
Every additional nine cuts the permitted downtime by a factor of ten, while the cost of achieving it — redundancy, automation, testing — tends to grow faster than that.
Which period should I measure against?
The one in your agreement. A monthly target is stricter than a yearly one at the same percentage, because a single long outage cannot be averaged out over twelve months.
Can I add a target that is not in the list?
Yes. Type any percentages into the field, separated by commas, and the table rebuilds itself. The copy buttons export exactly what you see.