Sizing from measurements
We log what your current servers actually do and specify from that. A database starved for I/O gets NVMe, in-memory reporting gets RAM, and an old program that uses one thread gets fewer cores with a higher clock.
Not every workload tolerates company. A warehouse system in the Koper hinterland may need to print hundreds of picking lists before the six o'clock departures; an automotive supplier in Novo mesto may rely on its ERP to rerun material requirements every night. On a shared host, those jobs also depend on what strangers on the same hardware are doing at that moment. One month-end the close takes twenty minutes, the next it crawls on for two hours, and nobody can explain the gap. Leasing a dedicated server ends the guessing. You pay monthly for an entire physical box whose processors, memory and drives work for you and nobody else. Nothing is purchased, nothing new is depreciated, and no one has to walk into a server room. The facility looks after electricity, cooling and swapping broken components. We look after the layers on top, remotely: operating system, databases, virtual machines, backups and security. Because that division is written into the agreement, roles are clear long before anything fails.
Projects of this kind usually stumble when someone orders the beefiest model on the list. We prefer to measure, to count licences in from the start and to run the box by written routine.
We log what your current servers actually do and specify from that. A database starved for I/O gets NVMe, in-memory reporting gets RAM, and an old program that uses one thread gets fewer cores with a higher clock.
A data centre inside the EU with doubled power and uplinks, hardware swap times stated in the contract, and a processor agreement covering how the host treats personal data.
A RAID scheme suited to the job and out-of-band management, so we can watch the machine boot even when the OS is broken. Recovery does not have to wait for a person in the building.
Proxmox or Hyper-V can split the box so that ERP, file shares and remote desktops each run in their own VM, with separate resources and separate maintenance windows.
Firewall at the edge, VPN for other sites and home users, and admin traffic kept on a different segment from application traffic. Only services you have approved are visible.
A daily backup to deletion-proof storage elsewhere. If one hour offline would cost you more than a second lease, we add a warm spare that mirrors data continuously.
Patches, drive and RAID monitoring, fault tickets to the facility and checks after each repair, plus an annual review of size against price.
Delivery speed is set mostly by the host's stock. We commit to a switch-over date only after one full practice migration.
A week of CPU, memory, disk and network logging on your existing systems, ideally spanning a month-end or a stock count.
One or two machines priced with lease, licences and management, next to the same sums for a large VPS, so the decision is arithmetic.
After installation we migrate once against the clock. That measurement defines the real cut-over slot, always outside office hours.
The switch, sign-off from the departments that lean on it most, then routine operation with monitoring and patch dates.
Watch the licence bill, not just the lease. Windows Server and SQL Server are typically licensed per core. A machine with twice the cores might add only a little to the monthly lease, yet push Microsoft costs up dramatically even though the ERP never uses those cores. So lease and licences are always added up together, and very often we end up proposing fewer, quicker cores.
Off-the-shelf configurations often ship within days; bespoke builds can take weeks. From order to live service, practice migration included, two to four weeks is a fair estimate.
An alert fires, we open a ticket with the facility, their technician swaps the drive and we follow the rebuild, all while the server stays up. Remember that RAID is about hardware only; a mistakenly deleted folder or a ransomware attack is handled from the off-site copy.
That is colocation, and it needs racking, cabling and hands-on hardware work that we do not provide, so a local contractor would be involved. With a leased machine the hardware belongs to the facility, which removes that need.
We move everything to its new destination, wipe the old machine and keep a record of the wipe. How the host disposes of drives afterwards, and whether it confirms that in writing, is settled at the start; the confirmation then sits with your retention and erasure documentation.
Monday to Friday, 9:00-17:00 Central European Time, with response times by plan: 4 hours for Start, 1 hour for Business, 15 minutes for Premium. Premium clients also get a 24/7 duty line written into their contract.
Which systems will it carry, how big are the databases, and does data have to stay in a particular country? We will propose suitable machines with all costs included.
Enquiry received
An engineer will write back no later than the following working day. Faults that leave your staff unable to work are always looked at first.
No town by that name here. Check the letters, or settle for the nearest larger place. Our engineers work only over the internet, so Kočevje and Koper get precisely the same treatment.