
NIS2 is the European Union’s cybersecurity law for essential services, and it is now in force. If your organization falls under it, you have probably read that you must “keep logs” without finding a clear statement of what to log or how long to keep it. That is because the answer is split across two legal texts, and much of what circulates online fills the gap with numbers the law never states.
This guide sets out the NIS2 logging requirements from the official texts: the NIS2 Directive itself and the Commission Implementing Regulation that adds the technical detail. We will separate what the law actually says from what people assume it says. We build UTMStack, a SIEM (security information and event management) tool, so weigh our view accordingly; most of this guide is about the law, not our product.
The short answer: the NIS2 Directive, Directive (EU) 2022/2555, never uses the word “logging.” It requires incident handling and ways to check that your security measures work, and logging is how you do both. The detailed logging rules live in a separate text, Commission Implementing Regulation (EU) 2024/2690, and they bind only certain digital providers such as cloud, data centre, DNS and managed service companies. Those rules tell you to keep, review and back up logs of named activities, and to run monitoring continuously, but they set no fixed retention period. You choose a “predefined period.” On top of that, every entity must report a significant incident within 24 hours and 72 hours, which is impossible without logs you actually watch.
Two levels of NIS2 logging duty
Where the rule that binds you comes from depends on what kind of entity you are
Directive (EU) 2022/2555, Article 21(2): incident handling and checking that security measures work. Logging is not named, but it is how you meet these. The detail is set by your country's own NIS2 law.
Regulation (EU) 2024/2690, Annex point 3.2: cloud, data centre, DNS, CDN, managed service and managed security providers, marketplaces, search engines, social networks and trust services get a written list of monitoring and logging duties.
Where the NIS2 logging rules actually live
NIS2 replaced the older 2016 network and information security directive. European Union member states had to turn it into national law by 17 October 2024. The Directive sets out risk-management measures in Article 21(2). None of them says “log,” but two make logging unavoidable:
- Point (b), “incident handling.” You cannot handle an incident you cannot see, and you cannot see one without records of what happened.
- Point (f), “policies and procedures to assess the effectiveness of cybersecurity risk-management measures.” Checking that a control works means reviewing evidence, which means logs.
The explicit, written logging rules are in a second text. Commission Implementing Regulation (EU) 2024/2690 was adopted on 17 October 2024 and entered into force on 7 November 2024. It is directly applicable, meaning it binds entities without each country rewriting it. Its Annex, point 3.2 “Monitoring and logging,” is the closest thing NIS2 has to a checklist. The catch is scope: the Regulation applies to the digital providers named in its Article 1, not to every NIS2 sector.
| Source | What it covers | Who it binds |
|---|---|---|
| Directive (EU) 2022/2555, Article 21(2) | The ten risk-management measures, including incident handling and effectiveness checks | Every essential and important entity, through national law |
| Regulation (EU) 2024/2690, Annex point 3.2 | A written list of monitoring and logging duties | The digital providers in the Regulation, such as cloud, data centre, DNS and managed service companies |
| Your national NIS2 law | Country-specific detail, sometimes stricter | Every entity in that country |
So if you run a hospital or an energy grid, the binding text is Article 21(2) as written into your country’s law, read with guidance from ENISA, the European Union Agency for Cybersecurity. If you are a cloud or managed service provider, the Regulation’s Annex applies to you directly. Many online guides quote the Regulation’s Annex as if it binds everyone. It does not.
What NIS2 says to log
For the digital providers covered by Regulation (EU) 2024/2690, Annex point 3.2.3 says to maintain, document and review logs. It does not hand you a list of event IDs. Instead it says to base the list on risk and, where relevant, to cover a set of activities. Those activities map cleanly onto the kinds of events a SIEM collects.
| What the Regulation points to | What it means in practice | What it catches |
|---|---|---|
| Network traffic | Firewall, flow and proxy records | Scanning, unexpected connections, data leaving the network |
| Account changes | New users, group membership, deletions | Backdoor accounts and quiet privilege grants |
| System access and authentication events | Logins, failures, multi-factor prompts | Brute force, password spraying, stolen credentials |
| Privileged activity | Admin and root actions | Misuse of powerful accounts |
| Configuration and backup-file changes | Edits to settings and to backups | Tampering meant to weaken defences or destroy recovery |
| Physical access | Door and rack access records | Someone reaching the hardware directly |
Point 3.2.1 requires procedures and tools to monitor and log activity so incidents can be detected and handled. Point 3.2.2 says this monitoring must be automated where feasible, and run continuously or at regular intervals. Point 3.2.4 requires regular log review and thresholds that raise an alarm. Point 3.2.6 requires a synchronised time source across systems and redundant storage of logs. Taken together, these describe central log collection with automated detection, which is what a SIEM does.
The retention myth
The single most repeated claim about NIS2 is that it sets a log retention period, often quoted as six or twelve months. Checked against the official texts, this is false.
Neither the Directive nor Regulation (EU) 2024/2690 names a number. The only wording is in Annex point 3.2.5, which says to maintain and back up logs “for a predefined period” and to protect them from unauthorised access or change. The period is yours to define, based on risk, investigation needs and any other law you are under. A figure you read on a vendor page is that vendor’s suggestion, not the text of NIS2.
This matters in two directions. You cannot point to NIS2 to justify deleting logs early, because it does not permit that. And you cannot assume a fixed number satisfies it, because it asks you to set and document your own period. In practice, many entities land on twelve months or more because other rules, such as those for finance or data protection, push them there, and because incidents are often found long after they start.
Why the 24 and 72 hour clocks need logs
NIS2 incident reporting has three deadlines in Article 23, and all of them depend on records you can search fast.
The NIS2 incident reporting clock
Each stage runs from when you become aware of a significant incident
- Within 24 hoursEarly warning
Tell the authority an incident is under way, and whether it looks malicious or could spread across borders.
- Within 72 hoursIncident notification
Give an assessment: severity, impact, and signs of compromise where known.
- Within 1 monthFinal report
A full account: root cause, the steps taken, and the cross-border effect.
A 24-hour early warning means you need to know, on the first day, that something is wrong and whether it is likely an attack. A 72-hour notification means you need enough detail to assess severity and impact. Neither is realistic if your logs are scattered across machines, missing for the hosts that matter, or too slow to search. This is the real reason NIS2 pushes central logging, even though it never says “retain for X months.”
How UTMStack fits, and where it stops
To be clear first about a limit: UTMStack does not ship a built-in NIS2 report or control mapping. We maintain mappings for nine frameworks, including ISO 27001, NIST and GDPR, and NIS2 is not yet one of them. So UTMStack will not tick NIS2 boxes for you. What it does is the underlying work the Regulation’s monitoring-and-logging points describe:
- Central collection with automated detection. UTMStack gathers logs from servers, firewalls, identity systems and cloud accounts, and checks each event against more than 600 detection rules mapped to the MITRE ATT&CK catalogue of attacker techniques as the event arrives, before it is stored. That matches point 3.2.1 and the “automated” wording of 3.2.2, and it is what gives you a chance at the 24-hour clock. See what logs should be sent to a SIEM for how to choose sources.
- Retention you define. UTMStack keeps events for up to five years with searchable cold storage, so you can set a “predefined period” under point 3.2.5 rather than being forced into one. Logs held off the original host also survive an attacker wiping that host, which supports the redundancy in point 3.2.6.
- Review and alerting. Dashboards and alerts cover the regular review and alarm thresholds in point 3.2.4.
- European data residency. For entities that must keep data in the European Union, UTMStack Cloud runs as a dedicated instance in France, Poland or the United Kingdom, among other regions. If you would rather hold everything yourself, UTMStack is self-hosted and installs on your own Linux servers, with an air-gapped option.
Two honest cautions. First, a synchronised time source under point 3.2.6 is something you set on your systems, usually with a network time service; a SIEM relies on it but does not provide it. Second, no tool makes you compliant on its own. NIS2 is mostly about governance, roles and process, and logging is one measure among ten. For the wider picture, see our guide to SIEM and compliance and the compliance overview.
Frequently asked questions
Does NIS2 require a SIEM? No. NIS2 does not name any product or product category. It requires incident handling, monitoring and the ability to detect and report incidents quickly. A SIEM is the common way to meet those requirements, because central logging with automated detection is hard to do any other way at scale, but you are free to meet the duty differently.
How long must I keep logs under NIS2? There is no fixed period in the law. Regulation (EU) 2024/2690 says to keep and back up logs for a “predefined period” that you set based on risk. Many organizations choose twelve months or longer, often because other rules require it and because incidents are frequently discovered late, but that is a choice, not a NIS2 number.
Do the detailed logging rules in Regulation 2024/2690 apply to my company? Only if you are one of the digital providers it names, such as a cloud, data centre, DNS, content delivery, managed service or managed security provider, an online marketplace, a search engine, a social network or a trust service. Other sectors follow Article 21(2) of the Directive as written into their national law, which does not spell out logging in the same detail.
What exactly does NIS2 say to log? The Regulation’s Annex points to network traffic, account changes, system access and authentication, privileged activity, configuration and backup-file changes, and physical access, chosen on a risk basis. It does not give event IDs. You decide the exact sources, document the choice, and review the logs regularly.
When did NIS2 start applying? Member states had to transpose the Directive by 17 October 2024. Regulation (EU) 2024/2690 entered into force on 7 November 2024 and applies directly. Certain digital providers also had to register with their authorities by 17 January 2025.
Sources
- Directive (EU) 2022/2555 (NIS2): https://eur-lex.europa.eu/eli/dir/2022/2555/oj
- Commission Implementing Regulation (EU) 2024/2690: https://eur-lex.europa.eu/eli/reg_impl/2024/2690/oj
- ENISA, technical implementation guidance on the cybersecurity risk-management measures: https://www.enisa.europa.eu/
- MITRE ATT&CK, Enterprise techniques: https://attack.mitre.org/techniques/enterprise/


