This Service level agreement is CheapServ's commitment on the availability of the infrastructure you rent: network and power in each of our six locations, and the hardware under every dedicated server. It forms part of the Terms of service. Service credits under this agreement are added to your prepaid balance.
1. Our commitment
- Network availability
- 99.99% per calendar month, per location: your server's port can send and receive traffic to the public Internet through our uplinks.
- Power availability
- 99.99% per calendar month: your server receives power from the rack distribution.
- Hardware replacement
- On dedicated servers, a failed disk, memory module or power supply is replaced within 4 hours of diagnosis, from the spare pool on site.
- Scheduled maintenance
- Announced at least 72 hours in advance on /status and by e-mail.
99.99% over a 30-day month allows for about 4 minutes and 20 seconds of downtime. Each location has two 100 Gbps uplinks to independent carriers and peering at the local exchange, so a single carrier failure does not count as downtime for you.
2. Definitions
- Downtime is a period of at least 60 consecutive seconds during which your server cannot exchange traffic with the Internet because of a failure of our network or our power distribution, or, on dedicated servers, because of a hardware failure we are responsible for. Downtime is measured from the moment our monitoring detects the failure, or from the moment you open a ticket, whichever comes first, until service is restored.
- Availability for a month is: (minutes in the month minus downtime minutes) divided by minutes in the month, expressed as a percentage.
- Monthly fee is the price of the affected server for the month in which the downtime occurred, as shown on its invoice.
- Service credit is an amount in USD added to your prepaid balance. It can be spent on any CheapServ service.
3. Service credits
If availability in a calendar month falls below 99.99%, you are entitled to a credit calculated on the monthly fee of each affected server:
| Monthly availability | Downtime in a 30-day month | Credit |
|---|---|---|
| 99.9% to 99.99% | 4 min 20 s to 43 min | 10% of the monthly fee |
| 99.0% to 99.9% | 43 min to 7 h 12 min | 25% of the monthly fee |
| 95.0% to 99.0% | 7 h 12 min to 36 h | 50% of the monthly fee |
| Below 95.0% | More than 36 h | 100% of the monthly fee |
Credits are calculated per server and per month. Several incidents in the same month are added together before the table is applied.
4. How to claim
- Open a ticket at /support within 7 days of the end of the incident. Include the hostname or IP address of the affected server and the start and end times in UTC as you observed them.
- We compare your report with our monitoring and the incident log published on /status. If our records show a shorter downtime, we tell you what we measured and how.
- Once verified, the credit is added to your balance, normally within 5 business days, and appears in your invoice history.
Incidents listed on /status as affecting your location are credited on request without further evidence: quoting the incident and your server is enough.
5. Hardware replacement
Every location keeps a spare pool of disks, memory, power supplies, network cards and complete chassis for the models we deploy. On dedicated servers:
- A failed disk, memory module or power supply is replaced within 4 hours of diagnosis. Diagnosis starts when our monitoring raises an alert or when you open a ticket.
- When a full chassis swap is needed, we move your disks to the replacement chassis whenever the platform allows it, so your data and configuration survive.
- Time beyond the 4-hour window counts as downtime under section 2. Data on a failed disk is not covered: keep backups, or add the optional backup service on VPS.
On VPS, GPU servers and RDP, a host failure is handled by us: your server is restarted on healthy hardware and the interruption counts as downtime.
6. Scheduled maintenance
Maintenance that may interrupt service is announced at least 72 hours in advance on /status and by e-mail to the owners of affected servers, with the expected duration. We schedule it in the low-traffic window of the location. Announced maintenance does not count as downtime as long as it stays within the announced window; any overrun does count.
Emergency maintenance to fix an active security issue or an imminent hardware failure may be performed with shorter notice. It is posted on /status as it happens and is credited as downtime if it exceeds 15 minutes.
7. Exclusions
The following do not count as downtime and do not give rise to credits:
- Scheduled maintenance announced under section 6.
- Outages caused by your configuration, your software, your operating system, a full disk, or actions taken through your account or API keys.
- Suspension of a server for non-payment or for a breach of the Acceptable use policy.
- Denial-of-service attacks exceeding the capacity of the included protection, and the null-routing of an address to protect the rest of the network during such an attack.
- Failures outside our network: your ISP, transit providers beyond our uplinks, DNS providers, or Internet routing incidents that do not originate from us.
- Force majeure: fire, flood, extended utility failure at the facility, and other events beyond our reasonable control.
8. Limits
Credits are your sole remedy for downtime. The total credit for a server in a month cannot exceed 100% of its monthly fee. Credits are added to your balance, are not paid out in cash or cryptocurrency, and cannot be transferred to another account. A claim must be made within 7 days; later claims are not accepted. Repeated failure to meet this agreement in three consecutive months for the same server entitles you to cancel that server and receive a refund of its unused balance under the Refund policy.
9. Changes to this agreement
We may revise this agreement when our infrastructure changes. The effective date at the top of the page is updated with every revision. Changes that reduce a commitment are announced on /status and by e-mail at least 30 days before they apply, and never apply to a month that has already started.
10. Contact
Claims and questions about this agreement: /support or [email protected]. Live component status and incident history: /status.
Something missing or wrong on this page? Open a ticket and tell us.