What Article 21 actually says
Article 21(1) requires an all-hazards approach aimed at protecting network and information systems and the physical environment of those systems from incidents. This extends analysis beyond malware and intrusion to events that interrupt the systems' availability or safe operation.
An all-hazards review still has a defined focus: the risks to systems used for the entity's operations or provision of services. It does not turn NIS2 into a general obligation to eliminate every business risk.
Threat scenarios that cross physical and digital boundaries
Consider a water utility whose monitoring network depends on a communications link, local power and a supplier-operated platform. A network intrusion might corrupt telemetry, while flooding could damage a control room and a power failure could disconnect sensors. Different initiating events can produce the same loss of visibility.
Risk analysis should identify shared points of failure. If the backup site shares the same power feed or supplier as the primary site, a seemingly separate recovery measure may not provide independence.
How to carry out an all-hazards assessment
Begin with the service and the systems that sustain it. Identify threats to confidentiality, integrity and availability, then consider operational and physical dependencies. Assess likelihood and plausible consequences, including duration of service disruption and knock-on effects for users.
Document existing safeguards against each material scenario. These may combine physical access controls, redundancy, secure configurations, incident response, maintenance arrangements, backup restoration and communication procedures. Assign an accountable person to close important gaps.
How proportionality changes the response
NIS2 does not require every entity to deploy the same physical protections or cyber safeguards. The relevant response depends on exposure, size, likelihood, severity and possible societal and economic consequences. A remote service operator with a single critical facility may need a different continuity design from a geographically distributed provider.
Document decisions where a threat is considered low priority, particularly if the event could interrupt an essential service. Assumptions about recovery capability should be tested rather than accepted only on paper.
Connect assessment findings to controls and evidence
Risk scenarios are useful only when they influence action. Link each significant exposure to prevention, detection, response or recovery arrangements. Retain system diagrams, dependency registers, risk records, exercise outcomes and management decisions so the assessment can be repeated after infrastructure or supplier changes.
Points of clarification
Does all-hazards mean only cyberattacks?
No. It also includes relevant accidental, physical and environmental events that threaten the systems and their supporting physical environment.
Is a separate disaster recovery plan enough?
Not by itself. The organisation should integrate continuity planning with threat assessment, preventive measures, incident handling and evidence of testing.