Article · 6 min read

Who reads the logs in your plant?

A recent public case shows what happens when each security tool keeps its own logs and nobody reads them together. This article gives you a first step of one hour and a one-page routine for a small team.

What the minister said

Ivan Vasilev, Bulgaria's Minister of Innovation and Digital Transformation, described a long cyber-espionage operation against the state administration. According to the minister, international hacker groups had access to state systems for almost five years, from 2021. They took at least 130 GB of information.

The attackers got into the infrastructure that institutions use to exchange information. The investigation found that the attacks reached servers, information systems, workstations and network devices in many institutions. The first signal came on 23 June 2026 from the Cybersecurity Centre of Information Services, the state IT company.

The minister gave one reason for the long delay. The security tools that were in use did not work as one integrated system. That made it difficult to see coordinated attacks across different institutions. In his words, the new national cybersecurity system "for the first time gives a single picture of what is happening in the state networks". The activity was found after that system went live.

According to the minister, the attackers' access is now cut, the affected systems are isolated and the malware is removed. The ministry also sent the case to the prosecutor. Who is behind the operation is not yet known.

We know only what the public reports say. We do not comment on the investigation. We look at one point from the briefing, because it applies to every plant.

Why this matters for a plant

A plant has protection in several places. A firewall sits between the office and the control network. A gateway or a VPN lets the integrator connect. SCADA and HMI servers log each login. Switches, servers and the alarm system each keep their own logs.

Each of these works on its own. The gap is between them. One person sees the firewall log. Another person sees the SCADA login log. Nobody compares the two.

Small operators feel this most. For example, a water company can have three automation engineers and two IT people, and nobody with "security" in the job title. Each side thinks that the other side reads the logs.

In our 15 years on the operator side of energy, the logs usually existed. Reading them was often nobody's named task.

Having logs and reading logs

Use this table to see where your own logs stand. Fill in the last column for your plant.

SourceWhat it can showOften held byWho reads it?
Remote access gateway or VPNWho connected, when, from where, for how longIT or the integrator
Firewall between the office and the control networkWhich connections passed and which were blockedIT
SCADA and HMI login logsWhich account logged in, and what changedAutomation
Windows servers and the domain controllerNew accounts, administrator use, failed loginsIT
Switches and other network devicesNew devices, configuration changesIT or the integrator
Alarm systemAlarms raised, silenced or changedOperations

An empty cell in the last column is a finding. It is a normal finding, and a name and a date can fix it.

Four things can stop a log from being read:

  1. The log stays on the device, and new entries overwrite old ones after a short time.
  2. The log goes to a server that nobody opens.
  3. IT reads the log, but does not know which control system login is normal.
  4. Operations knows what is normal, but cannot see the network side.

In the last two cases, each side has half of the answer. A short joint review joins the two halves.

A first step for next week: follow one remote access path

You need one hour with an IT person and an automation engineer. You do not need to buy anything.

  1. Choose one remote access path into your control network. The integrator's connection is a good choice.
  2. Find where each connection is logged. Write down the device and the place where the log is kept.
  3. Find out how long the log is kept.
  4. Find out who reads it, and when they last did. Write a name, not a team.
  5. Ask the shift supervisor and the integrator what a normal connection looks like: who, at what time, for how long.
  6. Write the answers on one page.

If an answer is missing, that is your first task. Give it an owner and a date. Do not choose a tool before you finish this page.

A one-page review routine

When you know where the logs are, set a short weekly review. Twenty minutes is enough for most small sites. The shift supervisor or an IT person can do it. Use one page with five checks:

CheckWhere to lookResult
Remote access outside the agreed hours, or by a person not on the listGateway or VPN log
New accounts, or administrator accounts used outside maintenanceServers and SCADA
New devices on the control networkSwitch and firewall logs
Alarms disabled or changedAlarm system
A log that stoppedEach source above

Write the result of each check in one line. Sign and date the page. Keep the pages in one folder.

The folder has a second use. When an inspector asks what you check and when, you can show it. Show the page to management too. Under NIS2, management approves and supervises the security measures.

Check where your alarms are

Some plants raise their critical alarms in the same system that an attacker can change. If a person with access can change a setpoint and silence the alarm in one place, the operator sees nothing wrong.

Test one alarm next week:

  1. Choose one alarm that protects the process or the product: a temperature, a level or a pressure.
  2. Find the system that raises it.
  3. Find out who can change or disable it in that system.
  4. Find out if a second, independent signal exists, for example a hardwired alarm, a separate device or a local indicator.

Then ask the shift one question: "If SCADA showed a wrong value, how would you know?" If the answer depends on SCADA itself, add the independent signal to your task list.

How this relates to NIS2

NIS2 Article 21 lists the security measures that organisations in scope must have. Three of them apply directly here: incident handling, a way to check that your measures work, and control of who has access to what. These measures apply to OT as well as to office IT.

Article 23 sets the reporting times. You must send an early warning within 24 hours after you become aware of a significant incident. You cannot report what nobody sees. A log that nobody reads makes that moment come later.

The Directive does not set a log format or a retention period. National law and sector rules can add detail. Check your national law, and ask your authority what it expects.

Do you need a monitoring tool?

Possibly, but not as the first step. A monitoring tool collects logs and raises alerts. A person must still read the alerts, and decide what to do at 02:00 on a Sunday.

Decide who that person is before you choose a tool. The one-page routine shows what you need to see and who can do the reading. If you need monitoring 24/7, a managed service is a better fit than a tool. We are not that service, but we can help you choose one.

Where to start

Do the one-hour path review next week. Then take theNIS2 readiness check for a first score in about one minute. If you want a second opinion on the result, a30-minute call can settle what to fix first.

Source: public statements by the Bulgarian Minister of Innovation and Digital Transformation, as reported in the Bulgarian media. The quote is our translation. This article is general guidance, not legal advice. Check the details against Directive (EU) 2022/2555 (NIS2) and your national law.

Not sure where you stand on cybersecurity and regulations?

Book a 30-minute call. Tell us about your operations. We tell you if we can help, and when we cannot.