Cloud Servers

Cloud Servers offer on-demand availability of computer system resources, including data storage and computational capacity without the need to directly manage the server hardware. UpCloud users are able to deploy Cloud Servers by selecting from the predefined configuration options; General Purpose, High CPU and High Memory plans or by freely allocating resources as needed using the Flexible plan.

General Purpose plans

General Purpose plans offer predefined configuration options with fixed monthly costs. Each plan bundles a set amount of server resources included in the plan pricing.

General Purpose plans include balanced sets of allocated CPU cores, amount of RAM, MaxIOPS storage and a monthly quota of outbound network transfer suitable for most use cases.

High CPU plans

High CPU plans offer sets of resources ideal for higher computational needs while being price competitive.

High CPU plans include the following configuration options for CPU cores, amount of RAM, MaxIOPS storage and a monthly quota of outbound network transfer.

High Memory plans

High Memory plans provide an increased amount of system memory for memory-intensive workloads.

High Memory plans allocated CPU cores, amount of RAM, MaxIOPS storage and a monthly quota of outbound network transfer according to the list below.

Predefined plans common features

The predefined configuration; General Purpose, High CPU and High Memory plans include a versatile range of server resources for varying use cases. While the resource configurations differ between each plan type, they all enjoy many of the same benefits.

Billing per hour with a fixed monthly price

Predefined plan Cloud Servers are priced by fixed monthly costs but billed per hour. This allows users to run persistent cloud servers at a predictable monthly cost but also deploy and delete short-term instances only paying for the hours used.

Cloud Servers on the Simple plan are billed hourly and capped at 672 hours (24 h x 28 days) per month. This is done to ensure the monthly costs are the same every month regardless of the number of days in each month.

What is included

The predefined plans consolidate everything needed to run many types of web applications with no obligations for additional services. The monthly price includes the CPU, system memory, and MaxIOPS storage listed on the pricing table, as well as 1 public IPv4 address.

What are billed separately

Besides the essentials, we also offer additional services such as backups, firewalls, floating IPs, and SDN private networking. These additional services and usages exceeding plan allowances, such as extra storage and network transfer exceeding the monthly quota, are billed hourly for their usage according to our listed pricing.

How predefined Cloud Servers plans are billed when shut down

General Purpose, High CPU and High Memory plan Cloud Servers are billed per hour regardless of whether the server is powered on or shut down. This is due to the discounted bundle pricing which in turn affords a lower monthly cost.

Trial limitations

We offer a 3-day free trial to new users which is intended to allow getting familiar with our services and test Cloud Server deployments without commitment. As such, the trial has some limitations. During the trial period, users are able to deploy only one €7 General Purpose plan Cloud Server at a time. However, it’s possible to delete and re-deploy a Cloud Server to test different locations or operating systems.

When the free trial expires, user-created resources are deleted. Users are able to continue using the service and retain all resources created during the trial by making at least the minimum €10 deposit to their account.

Flexible plan

In addition to the set configuration plans; General Purpose, High CPU and High Memory, UpCloud offers the option of selecting Cloud Server resources exactly as needed using the Flexible plan. The Flexible plan allows users to configure their Cloud Server with the number of CPU cores and the amount of system memory and storage as required. Each resource can be configured independently from the other allowing a wide range of Cloud Server setups.

Configuration ranges

By selecting the Flexible plan, users can adjust CPU and system memory resources on a fine-grain control. Flexible plans allow the following ranges of resource configurations:

• CPU 1 – 20 cores in 1-core increments
• Memory 1 – 128 GB in 1 GB increments
• Storage 10 GB – 4 TB in 1 GB increments, up to 8 storage devices

Billing per hour

Flexible plans are billed per hour according to the used resources without bundling. The estimated monthly cost is displayed in the UpCloud Control Panel in the server settings under the Resize menu. The resources are billed for every day of the month as opposed to the set plans which are only billed for 28 days per month.

How Flexible plan Cloud Servers are billed when shut down

Flexible plans also give users additional control over the costs of their cloud resources. Cloud Servers using Flexible plans are only billed for their storage, backups and network resources such as IP addresses and SDN private networks when powered down. When shutdown, CPU and RAM resources are not billed.

Resizing

Cloud Servers can be reconfigured even after deployment at the UpCloud Control Panel in the server settings and Resize menu. This allows users to increase or decrease their Cloud Server resources by simply powering down their server to make changes as required without the need for redeployment.

Scaling up

Increasing Cloud Server resources also called scaling up, allows users to allocate more CPU cores, system memory and storage when more capacity is needed.

Users can upgrade their Simple plan Cloud Servers easily by selecting a higher tier plan at the control panel in the server settings and the Resize menu.

Flexible plan users also have the option to increase Cloud Server resources to the same degree of control as at deployment.

Scaling up does not automatically increase the storage device size which needs to be done separately.

The Cloud Server must be in the stopped state during resizing.

Hot resize

Hot resizing allows the user to increase their Cloud Server resources without needing to shut down the server. The feature is automatically enabled on Cloud Servers created on or after April 27th 2022. Servers created before this date require a single shutdown and start for hot resizing to become available.

The maximum increase amount available is dependent on the user’s account resource limits and the available resources on the Cloud Server’s current host machine.

The storage devices attached to the Cloud Server being resized are not increased automatically. This is done to enable users to also easily scale down to their previous server plan. Increasing the storage device size needs to be done separately.

Hot resizing system memory is done using virtual memory slots. Each Cloud Server has a fixed number of memory slots. Every time memory is added using hot resized, these slots are used to include the added amount of memory. If all slots are already in use, system memory cannot be increased using hot resize until the server is shut down and started. After a full restart, allocated system memory is consolidated and the slots are freed for further hot resizing.

Scaling down

Both Simple and Flexible plan Cloud Servers can be scaled down by selecting a plan with fewer resources.

Scaling down does not reduce the storage device size automatically. This needs to be done manually by the user. This can be accomplished by following our storage resizing tutorial.

The Cloud Server must be in the stopped state during resizing.

Swapping between Simple and Flexible

In addition to being able to resize Cloud Servers, it’s possible to freely change between the Simple and Flexible plans at the UpCloud control panel under the resize menu. Selecting a plan of a different type will change how the Cloud Server is billed accordingly.

The Cloud Server must be in the stopped state during resizing.

Metadata service

UpCloud Metadata service allows users to query already deployed Cloud Servers to get information about the server itself. The data can be utilised, for example, as the basis for the automatic configuration of the servers after the initial deployment.

Enabling Metadata service

Metadata service can be selected at deployment or enabled on any existing Cloud Servers at the user’s UpCloud Control Panel or via the UpCloud API. The initial deployment of Cloud Servers using new public templates which require SSH-key only need to have Metadata service enabled.

Cloud Servers deployed without the Metadata service can enable it at any point without the need to restart the server. This can be done either via the UpCloud Control Panel in the server Optional settings or by using the UpCloud API.

The toggle for the Metadata service can be found in the Cloud Server settings under the Optional settings.

Alternatively, metadata service can be enabled with the following API request by replacing the server-UUID with the UUID of the user’s Cloud Server.

PUT /1.3/server/server-UUID/
{
   "server": {
      "metadata": "yes"
    }
}

Querying metadata

The Metadata is available as a cloud-init compatible JSON and via a filesystem-like traversable API. Both can be queried on any of the user’s Cloud Servers using any HTTP client such as curl or wget.

Metadata is available to be queried through all networks: public, utility and SDN private networks.

Metadata endpoint

http://169.254.169.254/metadata/v1.json

Metadata allows users to find specific information about resources which can include multiples of the same type, for example, storage and networks.

Below is a list of all currently available fields provided by the Metadata service.

Cloud name

"cloud_name": "upcloud"

The name of the cloud provider for inter compatibility

Instance ID

"instance_id": "00bf9504-a4cb-4839-88ff-124a2c95e169"

The UUID is generated at deployment

Hostname

"hostname": "metadata.example.com"

Cloud Server’s hostname as defined by the user

Platform

"platform": "servers"

The type of platform in use

Subplatform

"subplatform": "metadata (http://169.254.169.254)"

Metadata endpoint address

Public keys

"public_keys": [
  "ssh-rsa AAAAB[...]ud1Cw== [email protected]"
]

SSH keys included by the user

Region

"region": "de-fra1"

Server location

Network

"network": {
  "interfaces": [
    {
      "index": 1,
      "ip_addresses": [
        {
          "address": "94.237.90.209",
          "dhcp": true,
          "dns": [
            "94.237.127.9",
            "94.237.40.9"
          ],
          "family": "IPv4",
          "floating": false,
          "gateway": "94.237.90.1",
          "network": "94.237.90.0/24"
        }
      ],
      "mac": "de:ad:be:ef:3f:c5",
      "network_id": "03030473-8e9d-4f4f-bcfe-b2c300391a07",
      "type": "public"
    },
  ]
  "dns": [
    "94.237.127.9",
    "94.237.40.9"
  ]
}

Network interfaces and DNS addresses configured to the server

Storage

"storage": {
  "disks": [
    {
      "id": "0187b8c5-7220-4c90-9026-047dff8be0b3",
      "serial": "0187b8c572204c909026",
      "size": 25,
      "type": "disk",
      "tier": "maxiops"
    }
  ]
}

All storage devices attached to the server

Tags

"tags": [
  "dev",
  "metadata"
]

User-defined tags assigned to the Cloud Server

User data

"user_data": "apt-get update && apt-get -y upgrade"

Deployment script or script URL

Vendor data

"vendor_data": ""

Any specific data field can be queried using the filesystem-like traversable API. For example, the following query can be used to get the IP address of the first network interfaces.

http://169.254.169.254/metadata/v1/network/interfaces/1/ip_addresses/0/address

The Metadata service will then return the value of the queried field omitting the key. Meaning the output can be used directly without additional formatting.

94.237.90.209

The main advantage of Metadata allows it to be used to configure applications and services which require server-specific information that is not known before deployment.

Data centres

UpCloud runs data centres around the world to provide our users with a wide variety of regions and locations.

Our data centres are run in trusted enterprise-grade data centres that provide the state-of-art facilities needed to operate our services. We use our own hardware in every location.

Available locations

We currently offer data centres in the following locations:

Europe

• DE-FRA1 – Frankfurt, Germany
• FI-HEL1 – Helsinki, Finland
• FI-HEL2 – Helsinki, Finland
• ES-MAD1 – Madrid, Spain
• NL-AMS1 – Amsterdam, Netherlands
• PL-WAW1 – Warsaw, Poland
• UK-LON1 – London, UK

Asia-Pacific

• AU-SYD1 – Sydney, Australia
• SG-SIN1 – Singapore, Singapore

North America

• US-CHI1 – Chicago, USA
• US-NYC1 – New York City, USA
• US-SJO1 – San Jose, USA

New locations are added as we grow our infrastructure.

Different pricing in certain locations

Cloud Servers deployed to FI-HEL1 or FI-HEL2 are priced at 50% higher than our other locations. The prices are displayed at deployment.

Computation backend

UpCloud is built on the principles of speed and reliability by employing the latest technologies and hardware.

Automated failover

Our infrastructure is designed from the ground up with high availability in mind. All mission-critical components are backed up by N+1 redundancy, including all compute nodes hosting users’ Cloud Servers. Should a compute node fail, the operations will be transferred over to another compute node and the Cloud Servers will continue running with minimal interruption.

Virtualisation

We use the true and tested Linux Kernel-based Virtual Machine, or KVM, virtualisation module to provide hardware-assisted virtualisation.

CPU

We use top-of-the-range hardware at every one of our data centres, including high-powered enterprise-grade processors.

Automated balancing

Our infrastructure optimises the load across compute nodes within data centres by deploying new and recently started Cloud Servers on the least used hosts. Users experiencing reduced performance can take advantage of the automated balancing by shutting down and starting the affected Cloud Server.