When looking for a VPS server, you will encounter dozens of offers with similar-looking specifications and very different prices. Comparing VPS providers in the Czech Republic is complicated. You cannot rely solely on the figures in the offer. This guide explains how to compare correctly, what to watch out for, and what you should know before making your first payment.
Why comparing ventilation units is difficult
VPS specifications on paper are straightforward: X vCPU, Y GB RAM, Z GB disk, traffic, price. The problem is that behind the same figures there may be very different products:
- A vCPU is not a vCPU - it depends on the processor generation, how many virtual cores share a physical core (typically 2-8), and whether the vCPU is guaranteed or "burstable".
- NVMe is not NVMe - as discussed in the previous article, local NVMe and "NVMe-backed distributed storage" are very different in terms of performance.
- Traffic does not mean the same thing. - some count total incoming + outgoing traffic, some only outgoing, some not at all. "Unlimited traffic" may have a bandwidth cap (e.g. 100 Mbps), which is effectively a limit.
- SLA = server uptime, not your website's uptime - A 99.9% SLA covers the availability of physical infrastructure, not connectivity, DNS, or support response speed during an incident.
Key comparison criteria: What really matters
1. Computing performance: How to test
The most reliable way to compare computing performance is to run benchmarks after purchase or during the trial period. Synthetic benchmarks cited by providers in marketing materials are always measured under ideal conditions.
Practical tests:
- CPU benchmark:
sysbench cpu --cpu-max-prime=20000 run - Disk I/O:
fiowith a random read/write workload (see our article on NVMe vs SSD) - Network throughput:
iperf3 -c iperf.he.net - Latency to NIX.CZ:
ping nix.cz- under 5ms = server in the Czech Republic or a Prague data centre
2. Location and connectivity
Specifically, ask: In which data centre in the Czech Republic is the server physically located? Is the data centre connected to NIX.CZ and at what transmission capacity? How many BGP upstreams does the data centre have available?
Quality Czech data centres (O2 DC Prague, Wedos DC Písek, Cezar DC) have redundant connectivity and are direct members of NIX.CZ. Servers in such data centres offer low latency to the majority of Czech ISP customers and maintain good availability even if one peering link fails.
3. Virtualisation and resource guarantee
Ask: What is the virtualization technology? (KVM is standard, full hardware virtualization. OpenVZ/LXC are container technologies with limitations.) Are vCPUs guaranteed or shared? What is the overselling ratio?
An honest supplier will answer these questions specifically. Evasive answers or phrases like "our infrastructure is tailored to your needs" are warning signs.
4. Support: Speed and quality
The most critical aspect of support becomes apparent only during an incident, by then it is too late to find out. How to determine this in advance:
- Contact support before purchasing with a technical question (e.g., what the storage backend is) and measure the response time and quality of the answer.
- Check reviews on hosting comparison sites, specifically look for mentions of support during incidents, not just standard communication.
- Find out whether support is available 24/7 or only during working hours, and whether it is technical (able to resolve issues) or merely administrative (forwarding tickets).
5. Contract, terms and contractual penalties
For Czech-oriented projects, the advantage is a contract in Czech subject to Czech legislation. Specifically, review:
- Notice periods: can you cancel monthly or are you bound by annual payments?
- Conditions for a refund in the event of an outage exceeding the SLA
- What happens to your data when a service ends, how long are they obliged to retain it and in what format must they export it?
- Procedure in case of dispute escalation: arbitration or court proceedings?
Warning signs: What to watch out for
Overselling of computing resources
Overselling is the practice where a provider sells more virtual resources than it physically has available, based on the assumption that customers will not use them at 100% simultaneously. In reasonable amounts (a 2:1 ratio for CPU), this is standard. Aggressive overselling (10:1 or higher) means your VPS will be slow precisely when you need performance, during peak load, backups, or deployment.
How to check: Run a CPU benchmark at different times of day. If results fluctuate by more than 20%, overselling is likely.
Fake benchmarks and synthetic tests
Industrial synthetic benchmarks (Geekbench, PassMark, CrystalDiskMark) measured on a dedicated machine without other tenants are not transferable to shared VPS infrastructure. Always test yourself under production conditions.
Hidden fees
Typical hidden charges that only appear on the invoice:
- Price for IP addresses beyond the basic offer
- Overage charge for traffic exceeding the limit (can be expensive: read the tariff schedule)
- Charges for advance payment (if not included in the price)
- Setup fee upon order (a thing of the past in e-commerce, but still exists for hosting).
- Price for Windows licences or control panels (cPanel, Plesk): if you require these tools
Lack of transparency regarding infrastructure
A serious provider will specify the data centre, the hardware generation, the virtualisation technology and the storage backend. If you find only "modern infrastructure" on the website without further details and support fails to give a specific answer when asked. That is a problem.
Comparison criteria: Clear overview
Power output
CPU generation, NVMe type (local vs. Distributed), guaranteed IOPS, overselling ratio
Location
Specific data centre in the Czech Republic, NIX.CZ connectivity, latency, BGP redundancy
Price/performance
Price per vCPU/GB RAM, price per GB NVMe, hidden fees, overage rates
Support
Language (Czech), 24/7 availability, technical competence, SLA on response time
Terms and Conditions
Notice periods, SLA compensation, data export conditions, jurisdiction
Technology
KVM versus container, Proxmox / VMware, backup system, console for OS reinstallation
What Exascale.cz offers compared to standard competitors
We admit that we are a biased party in this comparison. But transparency is exactly what we expect from a hosting provider, and we should set an example.
Exascale.cz is a Czech infrastructure operated by Let IT Bee, s.r.o.:
- Virtualisation: Proxmox KVM, full hardware virtualization, not containers. Each VPS is an isolated machine with dedicated vCPU and RAM.
- Storage: Local NVMe disks on each host server: no distributed storage with network performance bottlenecks.
- Data centre: Czech data centre with connection to NIX.CZ: ping from the main part of the Czech Republic under 5 ms
- Overselling: CPU cores are not aggressively shared. We guarantee performance, not just allocation.
- Support: Technical support in Czech: not first-level scripts, but people who understand Linux administration.
- Transparency: Ask us anything about infrastructure. We will give specific answers.
This is not the cheapest option on the market, Hetzner in Frankfurt will always be cheaper. But for a project where Czech latency, Czech support and GDPR-friendly hosting in the EU matter, our price comparison makes sense.
Try Exascale.cz
Czech infrastructure, Proxmox KVM, local NVMe, technical support in Czech. Full root access within 60 seconds of ordering.
View price list →