Uptime & Availability

Loading…

← Back to dashboard · Storage health →

Server uptime

—

—

Availability (24h)

—

across all checks

Services

— up

—

Load average

—

—

Needs attention

Instance log

—

Probed every minute by log_uptime.sh. Availability counts a degraded service (unhealthy container, maintenance mode) as unavailable; checks with no data are excluded. 1h–30d are exact to the minute; 1y and all time are summed from a daily rollup, so they carry full history without keeping a year of per-minute rows. Time the box spent rebooting is reconstructed from its shutdown records and counted as downtime, so a reboot lowers these figures rather than disappearing from them.

Compute instances are not probed from outside: a machine its owner switched off and a machine that fell over look identical from here, so the EWS control plane is asked instead and reports which of the two it is. A deliberate stop counts as neither up nor down and is left out of the percentages, the same as any other unevaluated check.