UpCloud Cloud Servers and MaxIOPS Storage: A Practical Overview
A practical look at UpCloud cloud servers, covering plan families, MaxIOPS storage, networking options, the billing model and a sensible first deployment checklist.
UpCloud cloud servers are virtual machines from a European infrastructure provider headquartered in Helsinki, Finland, known mainly for fast block storage and a deliberately compact product range. This overview explains how the servers, the MaxIOPS storage layer, networking and billing fit together, so you can decide whether the platform suits your workload before you create anything.
Where UpCloud sits in the market #
UpCloud is an infrastructure-as-a-service provider rather than a hyperscaler. Its catalogue focuses on the building blocks most small and mid-sized teams actually use: cloud servers, block storage, private networking, managed databases, managed Kubernetes, object storage and load balancers. You will not find hundreds of niche services, and that is a design choice rather than an omission.
The company operates data centers across Europe, North America and Asia-Pacific, with a strong European footprint. For teams that care about where data is stored, being able to pick a specific European city for a server is often the first reason UpCloud makes the shortlist. For a broader summary of the company and its products, see our UpCloud platform overview.
How UpCloud cloud servers are organized #
A cloud server on UpCloud is a virtual machine with a fixed allocation of vCPUs and memory, one or more attached storage devices, and one or more network interfaces. You choose a plan, a zone (a single data center location), an operating system template and your storage size, then the platform provisions it through the control panel or the API.
Plan families
Plans are grouped by the balance of CPU and memory they offer. The exact names and sizes change over time, but the families generally map to familiar workload shapes:
| Plan type | Balance | Typical use |
|---|---|---|
| General Purpose | Even mix of CPU and RAM | Web apps, APIs, small databases, CI runners |
| High CPU | More vCPUs per GB of RAM | Build servers, batch processing, encoding |
| High Memory | More RAM per vCPU | In-memory caches, larger databases, analytics |
| Entry or developer plans | Small, low-cost sizes | Test environments, personal projects, staging |
Check the current plan list on UpCloud’s own site before sizing, because the provider adds and retires plan families. The principle stays the same: choose the family by the resource your application exhausts first, then pick the smallest size that leaves comfortable headroom.
Templates and images
UpCloud provides public templates for common Linux distributions and Windows Server. You can also turn a configured storage device into a private template and deploy new servers from it, which is a simple way to standardize a base image without adopting a full image-building pipeline on day one.
MaxIOPS: what it is and why it matters #
MaxIOPS is UpCloud’s own block storage technology and the default tier for server disks. Instead of relying only on local disks inside the physical host, MaxIOPS storage is network-attached and replicated, and it is built with a strong emphasis on input/output operations per second (IOPS). In plain terms, it aims to make disk reads and writes fast and consistent, which is where many small virtual servers struggle.
Why IOPS is often the real bottleneck
For many applications, CPU is rarely maxed out. What slows them down is waiting on disk: a database flushing its write-ahead log, a package manager unpacking thousands of small files, or a search index merging segments. These operations are dominated by small random reads and writes, and that is exactly the pattern IOPS measures. A storage layer tuned for high IOPS can make a modest server feel noticeably quicker than its vCPU count suggests.
Other storage tiers
Alongside MaxIOPS, UpCloud offers lower-cost tiers intended for data that does not need top performance, such as archives, backup targets or bulk files. Tier names have changed over the years, so confirm the current options in the official documentation. A sensible pattern is to keep the operating system and database on MaxIOPS and put large, rarely accessed data on a cheaper tier.
Backups and snapshots
UpCloud supports both manual storage backups (point-in-time copies of a storage device) and scheduled backups that you can enable per server with a retention policy. Backups are billed separately according to their size, so treat retention as a cost decision as well as a recovery decision. Always test a restore at least once; an untested backup is only a hope.
Networking on UpCloud #
Every server can have several network interfaces, each attached to a different kind of network. Understanding the three main types avoids accidentally exposing internal services.
- Public network: gives the server public IPv4 and IPv6 addresses reachable from the internet.
- Utility network: a private network that connects your own servers within your account, useful for backups and internal traffic.
- SDN private networks: software-defined networks you create yourself, with your own address ranges, so you can build a layout similar to a VPC with separate application and database segments.
UpCloud also offers floating IPs, which you can move between servers in the same zone for simple failover, and a firewall you can configure per server. Managed load balancers sit in front of groups of servers when you outgrow a single machine. If these concepts are new, our guide to VPCs, subnets, firewalls and load balancers explains them in provider-neutral terms.
How billing works #
UpCloud uses a pay-as-you-go model. Servers are billed by the hour while they exist, and each plan has a monthly price that acts as a cap, so a server that runs all month will not cost more than its listed monthly amount. Storage, backups, extra IP addresses and some network traffic are billed as separate line items.
Two details catch new users out. First, a stopped server may still incur charges for its storage and IP addresses, because those resources remain reserved for you; deleting the server and its storage is what stops the meter. Second, accounts typically work on a prepaid balance or a card on file, so keep an eye on the balance and any automatic top-up settings. Confirm the current rules on the official pricing and billing pages, as they can change.
A first deployment, step by step #
- Sign up on the official site at upcloud.com and complete the provider’s account verification and billing setup. New accounts may start with trial limits that are lifted once verification is complete.
- Enable two-factor authentication on the account and create separate sub-accounts for team members instead of sharing one login.
- Upload an SSH public key in the control panel so new Linux servers can be created without password logins.
- Choose a zone close to your users, or in the jurisdiction your data needs to stay in.
- Select a plan from the family that matches your workload, starting small. Resizing up later is generally straightforward.
- Pick MaxIOPS storage for the system disk and size it with room for logs and updates.
- Attach networks deliberately: public only where the server must be reachable, private networking for everything internal.
- Configure the firewall to allow SSH only from known addresses and open only the ports your application needs.
- Turn on scheduled backups with a retention period you can afford, then run a test restore.
Once a manual setup works, consider moving it into code. UpCloud has a public API and an official Terraform provider, which makes the same server, network and firewall layout reproducible. The UpCloud documentation covers the control panel, API and infrastructure-as-code tooling.
Where UpCloud fits, and where it may not #
Good fits
- Database-backed web applications where storage latency matters more than raw CPU.
- European teams that want to pick a specific country for data residency reasons.
- Small teams that prefer a short, understandable product list over a sprawling console.
- Workloads that need predictable monthly spend with the flexibility of hourly billing.
Things to weigh
- The managed service catalogue is much smaller than on AWS, Azure or Google Cloud. If you need a specific managed queue, data warehouse or machine learning service, you may have to run it yourself.
- Fewer regions than a hyperscaler, so check that a zone exists near your users.
- Third-party tooling, tutorials and community answers are less plentiful than for the largest platforms.
If you are comparing UpCloud with similar providers, our walkthrough of Hetzner Cloud servers, firewalls and snapshots covers a comparable European option, and the guide on object, block and file storage helps you decide which storage type each part of your application needs.
Frequently asked questions #
What is MaxIOPS on UpCloud?
MaxIOPS is UpCloud’s in-house, network-attached block storage technology and the default tier for server disks. It is designed for high input/output performance and redundancy, which particularly benefits databases and other disk-heavy workloads.
Does a stopped UpCloud server still cost money?
It can. Storage devices and reserved IP addresses usually continue to be billed while a server is stopped, because they remain allocated to you. Delete resources you no longer need, and check the official billing documentation for the current rules.
Can I manage UpCloud cloud servers with Terraform?
Yes. UpCloud publishes an API and an official Terraform provider, so servers, storage, networks and firewall rules can be defined as code and recreated consistently across environments.
Is UpCloud a good choice for beginners?
It can be, because the product range is compact and the control panel is straightforward. Beginners should still learn basic Linux administration, firewall configuration and backups, since an unmanaged cloud server leaves those responsibilities with you.
