Service status

Checking…

Fetching the latest measurements.

Last 24 hours: —
Last 7 days: —
Last 30 days: —
Last 90 days: —

Loading services…

Recent incidents

—

How this is measured

Every minute, from the front door

Each service is checked once a minute over its public address, with TLS, through the same gateway you use. A check that fails is retried once before it counts as an outage, so a single hiccup does not become an incident.

Every minute is kept

The result of every minute is stored for 90 days, so the percentages are counted, not estimated. That is why they can carry three decimals: 99.999% over 90 days is one minute of downtime, and it is one we actually recorded.

Slow still counts as answering

A service that answers, but slower than it should, is marked degraded. Those minutes count as available in the percentage and are shown separately in the bars, because they are not the same as being down.

Gaps are not outages

If the monitor itself is not running, those minutes are recorded as not measured and left out of the percentage. Downtime we did not see is never quietly turned into uptime.