Uptime commitment

Our commitment

We target 99.9% availability of hosting services in any calendar month. That works out at roughly 43 minutes of unavailability per month, and our intention is always to be comfortably below it.

“Availability” means our platform is able to serve your website over HTTP or HTTPS. It is measured across the platform, not from any single visitor's connection.

How we get there

Uptime isn't a promise, it's an architecture. The platform is built with:

• Redundant power, including UPS protection and on-site generation
• Multiple network transit providers with diverse paths and automatic re-routing
• Redundant storage, so a failed disk is a maintenance ticket rather than an outage
• Clustered web, database and mail services that tolerate a node dropping out
• An edge network that continues serving cached content independently of origin work
• Round-the-clock monitoring, so failures are handled rather than discovered

More detail on all of this is on our platform page.

What isn't counted

The following don't count towards unavailability:

• Scheduled maintenance, announced in advance where it may be disruptive
• Emergency maintenance required to preserve platform security or integrity
• Problems caused by your own code, applications, plugins, themes or configuration
• Accounts suspended for non-payment or for breaching the Acceptable Use Policy
• Issues with your domain registration, nameserver delegation or DNS hosted elsewhere
• Failures of your own connectivity or of intermediate networks between you and us
• Events outside our reasonable control, including large-scale attacks on third-party infrastructure

If we miss it

If we fall short of 99.9% in a calendar month, you can request a service credit against your next invoice, calculated pro-rata against the affected month's hosting fee.

• Email billing@rack1.net within 30 days of the end of the affected month
• Include your domain, your account details and the dates and times you observed the problem
• Credits are applied to future invoices and are the sole remedy under this commitment

We would obviously rather not owe you anything, and we'd rather tell you what went wrong than quietly issue a credit and hope you don't ask.

Knowing what's happening

When something is wrong at the platform level, we'd rather you heard it from us. Material incidents affecting hosting services are communicated by email to affected customers, with an explanation of what happened once it's resolved rather than only while it's ongoing.