Legal · Service level agreement
Uptime promises with consequences.
99.9% monthly uptime on network and host availability. On the Dedicated and Managed lines that commitment is backed by service credit; on Cloud VPS and the other virtual server lines it is the availability we target and report against, without a credit schedule attached. Where the credit schedule applies, you do not have to argue for it — only to claim it.
Version 1.1 · Effective 30 August 2026
- monthly uptime on network and host availability — at most about 43 minutes of downtime a month
- 99.9%monthly uptime on network and host availability — at most about 43 minutes of downtime a month
- maximum monthly credit when things go badly wrong, on the Dedicated and Managed lines
- 100%maximum monthly credit when things go badly wrong, on the Dedicated and Managed lines
- window to claim after an incident; credits applied within 7 days of the claim
- 14 dayswindow to claim after an incident; credits applied within 7 days of the claim
Credit schedule
What a miss costs us.
Credits are calculated against the affected service's charges for the calendar month and applied to your balance.
| Monthly uptime | Downtime (approx.) | Service credit |
|---|---|---|
| Below 99.9%, down to 99.5% | 44 min – 3.6 h | 10% |
| Below 99.5%, down to 99.0% | 3.6 h – 7.3 h | 25% |
| Below 99.0%, down to 95.0% | 7.3 h – 36.5 h | 50% |
| Below 95.0% | More than 36.5 h | 100% |
The maximum total credit in a month is 100% of the affected service’s monthly charge.
The terms
What counts, what does not, and how to claim.
What counts as downtime
A service is “down” when it is unreachable from the public internet because infrastructure under our control has failed: host hardware, the hypervisor, our network, or power and cooling in the facility. Downtime is measured from our monitoring or from your report, whichever starts earlier, and ends when connectivity is restored.
What does not count
- Scheduled maintenance announced at least 48 hours ahead and kept under 4 hours per month.
- Problems inside your operating system or applications — crashed processes, full disks, firewall lockouts. Support still helps, and the browser console and rescue mode exist for exactly this, but it is not platform downtime.
- Suspension for non-payment or for acceptable-use violations.
- Force majeure and third-party failures outside our and our upstreams’ reasonable control — DNS you host elsewhere, upstream scrubbing of a DDoS attack targeting you, submarine-cable cuts.
Claiming credits
Open a ticket titled “SLA credit” within 14 days of the incident, naming the affected service and the time window. We reconcile against monitoring data and apply the credit within 7 days. Credits offset future invoices; they are not refundable in cash.
Partner SLAs
Partners reselling PartVPS inherit this SLA for their end-customer workloads. Larger partners with committed volume can negotiate a stronger commitment — including a 99.99% target with reserved capacity — as part of a signed agreement. Start that conversation at [email protected] or via the partner programme page.
Put it to work
Reliability you can invoice against.
The commitment exists so you can pass it on — in your own customer contracts, with the credit schedule to back it.