All Systems Operational

How uptime is counted

• No downtime
• Minor downtime
• Major downtime
• Scheduled maintenance only
• No data

Every service is checked from inside our cluster every 30 to 60 seconds. A service counts as down when 2 or more checks in a row fail. A single failed check, or a response that arrived but was slower than its limit, is shown as degraded: it is listed for each day, but not counted as downtime. A day is green at 99.99% uptime or more, amber at 99.9% or more, and red below that. Days are UTC.

Scheduled maintenance: Sundays 02:00–05:00 UTC. Our servers install updates and restart one at a time in this window, and services can be briefly unavailable. Checks in the window are excluded from uptime; a day with nothing but maintenance is shown in blue.

Before per-check history was kept, only hourly totals were available. Days built from them count every failed check outside maintenance as downtime, and say so in their details.