Definition
Downtime is the time during which a website, API, or service is unavailable or failing to respond as expected. It is the opposite of uptime: if a service is up 99.9% of the month, the remaining 0.1% — about 43 minutes — is downtime.
Downtime isn't limited to a server being completely offline. A page that loads but returns an error, an API that times out, or a checkout that fails all count as downtime from the user's perspective, even if the host technically answered.
Why It Matters
Downtime directly costs money, customers, and trust. During an outage you lose sales, break integrations, and generate support tickets — and the damage is worse during peak traffic. Measuring downtime turns a vague sense of unreliability into a number you can reduce and report on.
How It Works
A monitoring tool checks your service at a set interval and records each failed check. Downtime is the total time those failures span. The precision depends on how often you check — a 1-minute interval pins downtime to the minute, while a 15-minute interval can only place it within a 15-minute window and may miss short outages entirely.
Real-World Example
An online store goes down at 2:03 PM and recovers at 2:25 PM — 22 minutes of downtime. With 1-minute checks, the monitor records the failure at 2:03 and recovery at 2:25, giving an accurate 22-minute outage. With 15-minute checks, it might only see the failure at 2:15, undercounting the real impact.
Best Practices
- Measure downtime with checks frequent enough to match the precision you need
- Treat error pages and timeouts as downtime, not just total server failure
- Track downtime per critical endpoint, not only the homepage
- Log every outage with start and end timestamps for later analysis
- Review recurring downtime patterns to find root causes
Common Mistakes
- Only counting full server outages and ignoring error responses
- Using infrequent checks that miss short outages
- Not recording downtime timestamps, making analysis impossible
- Measuring only the homepage while deeper pages fail unnoticed
- Reacting to downtime without tracking whether it's getting better or worse
In Monitoristic
Monitoristic detects downtime by checking your endpoint on a schedule and verifying the expected status code. When a check fails it opens an incident, sends Telegram and webhook alerts, and re-checks every 60 seconds until the service recovers.