Service · IT infrastructure

Performance tuning

Picture an architecture practice in Kranj with fifteen staff. Projects sit on a network drive in the office, and opening a large model takes several minutes. Anyone working from home avoids big files over the VPN altogether, copying them to a laptop in the morning and forgetting to put them back at night. On Wednesday afternoons, when everybody starts plotting drawings ahead of the client meeting, the network drive slows so badly that some applications crash. The usual first suggestion is a new server or faster disks. Before buying anything, it is worth finding out where the time actually goes. The answer is often surprisingly mundane: a switch running at 100 Mbit/s because of a damaged cable, antivirus scanning every file that opens, or a backup job that kicks off at two in the afternoon. We measure what happens during the slowest hours, pinpoint the bottleneck and remove it remotely. New hardware only comes up once free configuration changes are used up, and only if we can show the benefit in numbers. As we do not sell equipment, we have no reason to talk you into something bigger.

Peak-time measurement
not all-day averages
Settings first
purchases come last
Before and after
timed in seconds
Independent
no hardware to sell

Scope of the service

Slowness nearly always has one main cause and a few contributing ones. We work through the layers from the user's screen down to the database and web server.

Check the scope with an engineer

Workstations

Full disks, too little memory for the software in use, dozens of programs loading at start-up, and antivirus scanning a project folder that ought to be excluded.

File server or NAS

Disks rebuilding after a failure, a poorly configured SMB setup, folders holding hundreds of thousands of tiny files, and a server weighed down by backups in the middle of the working day.

Network

Ports negotiating the wrong speed, faulty cables, overloaded Wi-Fi access points, and traffic that should stay local travelling over the link between sites instead.

Database

Limits of the free SQL Server Express edition, missing or fragmented indexes, tables nobody has maintained, and a client version that does not match the server.

Remote access

Applications reaching an office database over VPN as though they were in the same room. Often it is better to run the program in the office or the cloud and send the user just the screen image.

Microsoft 365 and sync

OneDrive attempting to sync a hundred gigabytes of projects onto every laptop, and SharePoint libraries that have outgrown the recommended limits.

Online shop

Slow server response, uncompressed images, plugins firing dozens of queries per page, and stock synchronisation scheduled right at peak traffic.

How the work runs

Diagnosis takes one to two weeks. Fixes are rolled out gradually, in slots that cause the least disruption.

01

Slowness log

For a week staff note in a simple form what was slow and when. Precise timestamps are worth more than a general “everything is slow”.

02

Measurements

We time typical tasks such as opening a project or issuing an invoice, while an agent records load on the server, disks and network.

03

Diagnosis

The log and the measurements are overlaid. You get a list of causes with evidence and a recommendation on where to begin.

04

Fixes and re-measurement

Approved changes are applied remotely and the same measurements repeated. A month later we check the improvement has held.

The free edition has quiet limits. Plenty of business applications are installed on the free SQL Server Express in smaller deployments. That edition caps each database at 10 GB and restricts how much memory and processor power it can use. As the company grows, the database creeps towards the cap and the software slows down however powerful the server is. Checking database size and edition takes a few minutes and can spare you a purchase that would not have helped.

Your questions answered

Sometimes, though often not. If measurements show the server is constantly maxed out, you receive a technical specification and buy from a supplier of your choice. Physical installation is done by your staff or a local contractor, while we migrate data and configure everything remotely.

Diagnosis and fixes are billed at €50 per hour plus VAT against an estimate sent beforehand. We do not promise percentage gains before measuring, but after the fixes you see the difference in seconds.

Possibly in part. Measurements let the conversation with the vendor rest on data rather than guesswork. Changes to the application are made by the vendor; we look after the server, database and network.

Yes. We check the VPN, home broadband and how the application reaches its data. It often helps to run the program in the office or cloud and send only the screen home. Home-working rules such as equipment and expense reimbursement belong in a written arrangement under ZDR-1.

Yes. We measure Core Web Vitals and server response time, review caching, images and plugins, and show what your hosting cannot cope with. We only suggest switching provider if the measurements back it up.

Tell us what is slow

Describe one specific task that holds your team up and how long it takes. Within one working day we will suggest where to start measuring.

When we work
Weekdays 9:00-17:00 CET, answers the same or next working day
Calls
Video meetings, Teams and Meet both welcome

We set essential cookies only: the ones that keep the site running and recall which town you chose. Nothing for ads and nothing that follows you around the web. Details are in our privacy and data protection notice.