Service Level Agreement

In short

• We commit to 99.9% availability for every paid VPS, VDS, Minecraft, Discord bot and Lavalink service, every calendar month.
• For every 2 hours of downtime in a month, we add 1 free day to your service, up to 30 days a month.
• Within 15 days of a payment, if a fault on our side is not fixed, you can choose a calculated refund for the unused days instead, under our Refund Policy. It is never a full refund.
• Claim within 14 days of the end of the month, by ticket or email.

Scope

This agreement sets out the availability Rackyard commits to for every paid VPS, VDS, Minecraft, Discord bot and Lavalink service, and the compensation you receive if we miss it. It forms part of our Terms of Service and applies from the day a service is provisioned, for as long as it is active and paid.

Availability target

We commit to each service being available at least 99.9% of the time in each calendar month. That allows about 43 minutes of downtime in a 30-day month.

What counts as downtime

A service is down when, because of a fault in our hardware, network, power, virtualisation or game and bot hosting platform:

• it cannot reach the internet or be reached from it;
• it cannot be started, stopped or reinstalled from the control panel; or
• its disk cannot be read or written.

That includes a failure of the physical host your server runs on, even if no other customer is affected. Problems inside your server, such as its operating system, software or firewall, do not count, and neither does slower performance or packet loss on its own while the service remains reachable.

On Minecraft, Discord bot and Lavalink plans, we are responsible for the platform your server runs on, and you are responsible for what runs on it. A world, mod, plugin, bot or configuration that crashes, or a server that runs out of its plan’s memory, is not downtime. Lag is not downtime either, as long as the service is running and reachable.

Downtime is measured in minutes, per service, from when our monitoring detects the fault or you report it, whichever is earlier, until the service works again. All the downtime in a calendar month is added up, so several short outages count together. Outages with the same cause are one incident, but every minute of them still counts.

Compensation

For every full 2 hours of downtime a service has in a calendar month, we add 1 day to that service: its next due date moves one day later, at no charge.

• Days are added to the affected service only, whatever its billing cycle. They are not cash, cannot be moved to another service or account, and end if the service is cancelled.
• The most we add to one service for one calendar month is 30 days.
• Within 15 days of the payment for the current period, if a fault on our side stops the service working and we cannot fix it within 72 hours, you can choose to cancel and get a calculated refund for the unused days of that period instead. See our Refund and Cancellation Policy.

Added days are our remedy for missing the availability target. They do not limit rights you have under the Consumer Protection Act, 2019 or other law that cannot be excluded.

Data loss caused by us

If data on your service is lost or corrupted because of a verified fault on our infrastructure, such as a failed storage array, we will restore it from any backup we hold for you, including any backup product you bought from us, and tell you exactly what could and could not be recovered. Whether or not it can be restored, we add 30 days to the affected service. This does not apply where the loss followed from something done inside your server, or from suspension or termination under our Terms.

What is not covered

Time a service is unavailable because of the following does not count as downtime:

• planned maintenance announced with the notice below, and urgent security work, together up to 4 hours per service per month (anything beyond that counts);
• anything you or your users caused, including your own software, configuration or firewall rules;
• suspension under our Terms of Service, Billing Policy or Acceptable Use Policy, including for abuse or non-payment;
• denial-of-service attacks larger than our network can absorb, including the time a targeted address is null-routed to protect other customers;
• problems outside our network, such as your internet provider or a third party’s routing;
• outages of services we do not run but yours relies on, such as Discord, Minecraft sign-in, or the sites your audio comes from;
• events outside our reasonable control, as described in the Terms; and
• features we have labelled as beta or preview.

How to claim

Open a ticket in the client area, or email support@rackyard.cloud, within 14 days of the end of the calendar month in which the downtime happened. Tell us the affected service and roughly when it was down; you do not need your own monitoring to claim. We check the claim against our monitoring records and any evidence you send, and reply within 5 business days with the downtime we measured and the days we added. If you disagree with our figures, reply with your evidence and we will review it again.

For outages that are clearly ours and that we have announced, we aim to add the days without waiting for a claim, but a claim is how you make sure. Your account must have no overdue invoices when the days are added.

Incidents

When an incident affects customer services, we keep affected customers updated by email, from when we confirm the problem until it is resolved, with an estimate of when it will be fixed whenever we have one. For any incident that causes 2 hours or more of downtime, we publish an explanation within 5 business days of what happened and what we are changing so it does not happen again.

Planned maintenance

We schedule maintenance that may interrupt your service in advance and tell affected customers by email, at least 48 hours beforehand, with the expected length of the interruption. We do it at quiet times, usually between 01:00 and 06:00 local time at the location concerned. Urgent security fixes may occasionally need shorter notice; we tell you as soon as we can.