What Is Downtime?

Reviewed by Ionut Caval · Updated June 2026

Downtime is more than a site that fails to load. Depending on how you define "available", it can include a broken checkout, an expired TLS certificate, a DNS resolution failure, a 5xx error, or responses so slow they breach your performance threshold. What counts as downtime is a decision you make in advance, usually written into a monitoring rule or an SLA, so that every party agrees on when the clock starts and stops.

How downtime is measured

Downtime is the inverse of uptime: uptime percentage equals total time minus downtime, divided by total time, times 100. Because the percentages sound nearly perfect, the allowed downtime is easy to underestimate. A 99.9% target permits roughly 43 minutes of downtime per month, 99.95% permits about 22 minutes, and 99.99% permits only about 4.3 minutes. A single 30-minute outage can therefore consume most of a monthly 99.9% budget in one event. For a full breakdown, see what 99.9% uptime actually means.

The most common causes of downtime include:

• Infrastructure failures: a server, container, or database going offline.
• Network and DNS issues: routing problems or a misconfigured DNS record.
• Expired or misconfigured TLS certificates that block HTTPS connections.
• Deploys and configuration changes that introduce a regression.
• Traffic spikes or DDoS attacks that exhaust capacity.

Reducing downtime

Total downtime over a period is driven by two factors: how often failures happen and how long each one lasts. The second factor is MTTR (mean time to recovery), and the fastest way to lower it is usually to detect outages sooner, because you cannot start fixing a problem you do not yet know about. That is why teams measure downtime continuously and from multiple external locations rather than checking occasionally, and why uptime and SLA monitoring alerts the moment a check fails instead of waiting for a customer to report it.

See also: Uptime & SLA monitoring