Service level agreement
The availability commitment and credit policy for your VPS service.
1. Contractual availability
The instance and network availability commitment is 99.99% per calendar month. The account control plane and provisioning commitment is 99.9%. These are contractual percentages, not measured uptime results. The public status page does not currently publish a monitoring history.
2. Scope of downtime
Availability concerns whether the supplied instance is running and reachable, rather than whether your application is correctly configured. The published threshold is failure to respond to ICMP and a TCP connection on an available listening port from independent observation points for at least 60 consecutive seconds. An application failure or a guest firewall blocking intended access is not automatically a platform outage.
3. Credit bands
Credits are calculated per instance and calendar month against that instance’s monthly service charge.
| Monthly availability | Credit |
|---|---|
| At least 99.99% | None |
| 99.9% to below 99.99% | 10% |
| 99.5% to below 99.9% | 25% |
| 99% to below 99.5% | 50% |
| 95% to below 99% | 75% |
| Below 95% | 100% |
Credits are service credit, do not expire, and are not exchangeable for cryptocurrency. The published policy provides automatic eligible credits. If you believe an outage or credit has been missed, contact support with the instance reference and timestamps so it can be reviewed.
4. Exclusions
- Scheduled maintenance announced at least 72 hours in advance.
- Your own application, configuration or resource-exhaustion issues, including failure to unlock an encrypted volume.
- Suspension for non-payment or an AUP violation.
- An attack exceeding the applicable mitigation capacity.
- Force majeure such as natural disaster, war or state action. The existing policy provides a full refund for an affected service following a seizure rather than treating it as an SLA incident.
Provider hardware/software failures, carrier failures and facility power events are not excluded merely because they involve an upstream service.
5. Report an issue
Send [email protected] the order or instance reference, location, outage period and timezone, plus any independent observations. Do not include passwords or your full account key. No numeric latency, packet-loss or mitigation-speed benchmark is established by this page.