Definition
Uptime is the proportion of time that a website, API, or service is available and working as expected, usually expressed as a percentage over a defined period such as a month or a year. A service with 99.9% uptime was reachable and healthy 99.9% of the time and unavailable for the remaining 0.1%.
Uptime is the inverse of downtime: if a service is up 99.9% of the month, it was down for about 43 minutes. The figure only means something when paired with the period it covers and the way it was measured — a number measured with infrequent checks can look better than the reality users experienced.
Importantly, "up" should mean more than "the server answered." A page that loads but returns an error, or an API that responds slowly enough to be unusable, is not truly up. Good uptime measurement checks for a correct response, not just any response.
Why It Matters
Uptime is the single clearest signal of reliability. For a business, downtime means lost revenue, missed signups, broken integrations, and eroded trust — and the damage compounds during peak hours. Tracking uptime turns reliability from a vague feeling into a number you can measure, communicate, and improve. It is also the basis for any service level agreement (SLA) you offer or rely on.
How It Works
Uptime percentage is calculated as (time available ÷ total time) × 100. A monitoring tool checks the service at a regular interval, records each check as a success or failure, and aggregates the results over the reporting period. The accuracy of the number depends heavily on how often checks run: a 5-minute interval can miss short outages entirely, while a 1-minute interval captures far more, producing a more honest figure.
Real-World Example
A store is monitored every minute for 30 days (43,200 checks). During the month it has two outages totaling 22 minutes — 22 failed checks. Uptime = (43,200 − 22) ÷ 43,200 × 100 = 99.95%. Had the same site been checked only every 5 minutes, one of the short outages might have fallen between checks and never been recorded, inflating the reported number above the truth.
Best Practices
- Measure uptime with checks frequent enough to match the precision you care about — 1-2 minute intervals for revenue-critical services
- Define "up" as a correct response (expected status code), not just any response
- Report uptime alongside the period and check interval so the number is honest
- Track uptime per critical endpoint (checkout, API, auth), not just the homepage
- Pair uptime with response-time data to catch degradation before it becomes downtime
Common Mistakes
- Quoting an uptime percentage without saying over what period or how it was measured
- Using infrequent checks that miss short outages and overstate real availability
- Treating a 200 status as healthy when the page actually shows an error
- Monitoring only the homepage while deeper, more critical pages go unwatched
- Promising 100% uptime — no service can honestly guarantee it
In Monitoristic
Monitoristic checks your endpoints every 1, 2, or 5 minutes depending on your plan, records every check, and shows your uptime percentage with a 30-day history bar. Because it verifies the expected status code, "up" means genuinely working — not just reachable.