Linux server maintenance: what the contract covers

“The server works, what is there to maintain?” — that is the usual assumption until something breaks. In practice, Linux server maintenance is continuous work that happens in the background and whose result you only notice when it is left undone: the disk fills up, the certificate expires, a vulnerability goes unpatched or a backup will not restore.

In this article we look concretely at what proper server management covers, how to tell a serious maintenance contract from an empty promise and when it makes sense to buy management in at all.

What server maintenance covers

Updates and security patches

The foundation of everything. Linux distributions release security patches regularly and some of them are critical — a publicly known vulnerability is exploitable within hours of disclosure. Maintenance means someone watches those notices, assesses what affects this particular server and installs the updates so the services do not go down.

An important nuance: automatic updates alone are not maintenance. They solve the smaller part of the problem and occasionally create a new one when a package breaks a working configuration. Somebody has to check after an update that everything still works.

Backups — and testing that they restore

Having backups is half the job. The other half is checking whether they can actually be restored. Surprisingly often it turns out only in the moment of crisis that a copy was made partially, is corrupted or contains the database in an inconsistent state.

Proper maintenance means automatic copies kept outside the server and a periodic restore test. Without the test, a backup is only hope.

Monitoring and alerts

Disk usage, memory and processor load, whether services are alive, certificate expiry, unusual network traffic. The point of monitoring is that the problem is discovered before the customer reports it. Most failures announce themselves days in advance — a disk does not fill up overnight.

Security and access

Firewall rules, SSH access with a key only (not a password), blocking of attacks, review of user accounts and closing the access of employees who have left. The last one is the thing that is almost always forgotten.

Logs and performance

Watching the logs shows what is really happening on the server — which queries are slow, where the errors come from, whether someone is trying to break in. A performance review means keeping the database, web server and cache configuration in line with the load. A server that was correctly configured two years ago may no longer suit today's load.

Documentation and a recovery plan

Where the server is, what runs on it, who has access, how to restore. This is the part whose value becomes clear on the day the previous administrator is unreachable.

How to tell serious maintenance from an empty promise

When asking for an offer, ask four concrete questions:

1. How quickly do you respond and when? “Quickly” is not an answer. An answer is “within four hours on business days” or “within one hour around the clock” — and it has to be in the contract.
2. When did you last test restoring from a backup? If the answer is vague, the backups have not really been verified.
3. What is included in the price and what is extra work? Does a smaller change fit inside the monthly fee or does every one come with a separate invoice?
4. What happens when the cooperation ends? Do you get the access rights, the documentation and the backups? The answer must be an unconditional yes.

What server management costs

The monthly fee depends on the number of servers, the load and how fast a response is needed. Basic management of one ordinary web server is many times cheaper than a single serious outage during which an online store is down for half a day.

Our prices for Linux server setup and management are listed on the pricing page. If you already have a server that nobody manages, we will review it first and tell you honestly what state it is in — including when the conclusion is that no major work is needed.

When it makes sense to buy management in

If the server runs something whose downtime costs money — an online store, a customer portal, a booking system, internal company software — the answer is practically always yes. If it is a simple brochure site on shared hosting, separate server management is not needed; ordinary website maintenance is enough.

Another clear sign is a situation where only one person “knows” the server and nobody would know where to start if they left. That is a risk which always materialises at the worst possible moment.

Summary

Server maintenance is updates, verified backups, monitoring, security, performance and documentation — plus an agreed response time. Most serious failures can be prevented, and the ones that cannot are quickly solved when the preparation has been done.

If you would like someone to take care of it, get in touch — we will review your existing server and make a concrete offer.