服务等级协议

99.95% network and power availability per calendar month, per instance, measured at the hypervisor's uplink.

Fleet-wide availability over the last 24 months has been 99.987%. The commitment is deliberately lower than what we achieve, because a commitment should be one you can meet on a bad month.

Loss of network connectivity to the instance, loss of power to the host, or hypervisor failure, in each case lasting more than 60 consecutive seconds and confirmed by our own monitoring or by a ticket you opened.

• 01Scheduled maintenance announced at least 72 hours in advance, capped at 4 hours per quarter.
• 02Faults inside your instance: your kernel, your firewall, your full disk, your application.
• 03Suspension under the acceptable use policy.
• 04Loss of connectivity outside our network, including at your own ISP.
• 05Force majeure, and denial-of-service attacks exceeding the mitigation capacity published on the network page.

Downtime is credited at ten times its duration, as additional service days, up to 30 days per month.

Credits are calculated from our monitoring and applied automatically within five business days of the end of the month. You do not need to open a ticket, and asking for one does not increase it.

Median first response across all tickets was 11 minutes over the last quarter. Managed accounts carry a one-hour response commitment, 24/7.

Every ticket is answered by an engineer with hypervisor access. There is no first line and no script.