Software Growth

Uptime

Uptime is the percentage of time your service is available and working. 99.9% uptime allows about 8.8 hours of downtime a year.

Uptime tells customers how often your product works when they need it. It is usually quoted as a percentage over a month or a year, and it is the number most often written into a service level agreement. Engineers also call it availability.

How to calculate uptime

To turn a target into allowed downtime, flip it around:

Example: in a 30-day month (43,200 minutes), your product is down for 90 minutes across two incidents.

That misses a 99.9% target, which allowed only 43.2 minutes that month.

The nines

People describe availability in "nines". Google's SRE book notes that 99% is two nines and 99.999% is five. Downtime allowed per year (365 days, 8,760 hours) at common targets:

  • 99% (two nines): 87.6 hours, about 3.65 days.
  • 99.5%: 43.8 hours.
  • 99.9% (three nines): 8.76 hours.
  • 99.95%: 4.38 hours.
  • 99.99% (four nines): 52.6 minutes.
  • 99.999% (five nines): 5.26 minutes.

Per 30-day month, 99.9% is 43.2 minutes and 99.99% is 4.32 minutes. Each extra nine cuts the budget by 10 times, and the cost of reaching it rises faster than that: redundancy, multi-region setups, on-call rotations and slower, more careful releases.

What good looks like

EnterpriseReady says 99.9 percent is standard and 99.99 percent common in enterprise SLAs, and recommends a public status page showing real-time and historical availability. Google's SRE book argues against chasing perfection: its workbook states that 100% reliability is the wrong target, and that your error budget is 100% minus your SLO. With a 99.9% SLO you have a 0.1% budget to spend on shipping changes and absorbing failures.

Common mistakes

  • Measuring from your servers, not the user. A healthy server behind a broken login still means downtime for the customer.
  • Counting only full outages. Slow pages and failing checkouts count too if customers cannot work.
  • Excluding too much. Calling every incident "maintenance" fools nobody.
  • Promising a number you do not measure.

When a small SaaS should care

From day one, keep a simple external monitor and a public status page. Track your real number for a quarter before you promise anything. For most solo or small-team products, 99.9% is reachable on managed hosting and is enough for enterprise buyers. Unplanned downtime is a direct driver of customer churn, and ignoring reliability work builds technical debt that comes due during incidents.

Sources

Back to the SaaS glossary