Why Sentinel cannot serve an air-gapped network and what replaces it
Sentinel runs only in Azure, so it cannot be installed in a classified enclave, an OT network or any disconnected site where traffic to a cloud region is prohibited. Self-hosted Wazuh can run fully offline: its offline installation guide covers systems without internet connectivity, and its vulnerability detection can use an offline vulnerability repository in place of the Wazuh Cyber Threat Intelligence platform. The practical question for your team is therefore not whether a replacement exists but which components must be carried across the boundary and how they are verified on the way in.
The short answer is a Wazuh deployment with the Wazuh indexer, server, dashboard and agents all hosted inside the isolated network, with OpenSearch-based storage handled by the indexer rather than any external service. Wazuh's components and deployment options, including the offline installation guide, are described in its deployment options documentation. Confirm with Wazuh which release line and offline bundle match your accreditation scope before you start, because a 4.14 patch cadence means the offline update path will recur throughout the life of the system.
The Defender portal move after March 31, 2027 does not change the Azure dependency. Sentinel will be available only in the Microsoft Defender portal, which is a cloud service, so an enclave that could not reach Azure before still cannot reach it afterward. Plan the transition on the assumption that the air-gapped site needs a self-hosted platform regardless of Microsoft's portal changes. Ask your compliance lead which control evidence, such as supply chain verification of carried-in packages, must be retained in the offline environment.
What an air-gapped SIEM must satisfy
An isolated network changes what a SIEM needs in four places: installation, updates, vulnerability data and remote access. Each card below covers one of these requirements, with the controls and checks to confirm before you commit to a design.
No outbound internet at install time
Every package, container image, rule set and agent installer must be staged on media that crosses the boundary in a controlled way. Wazuh publishes an offline installation guide for systems without internet connectivity, so start there and confirm each component, including the indexer, server, dashboard and agents, installs from local sources. Ask the vendor which dependencies, such as operating system repositories, still reach external hosts.
Offline vulnerability data
Wazuh vulnerability detection correlates the software inventory collected from endpoints with vulnerability data from the Wazuh Cyber Threat Intelligence platform or an offline repository. An isolated site must use the offline repository, which means someone builds and refreshes it on an internet-connected system and moves it across. Agree on a refresh cadence and record the date of each import, because stale data produces false confidence in the findings.
Verified packages across the boundary
Every package, image and rule update carried across the isolated boundary needs a documented check of integrity and origin before it is installed. NERC CIP-013-2 addresses this through supply chain procurement processes that cover verification of vendor software and patches, where the site is in scope. Ask the vendor how release artifacts are signed, keep a transfer log with hashes, and define who approves each import.
Controlled vendor remote access
Where a site falls under NERC CIP-005-7, remote access for vendors to high and medium impact BES Cyber Systems must pass through an Intermediate System with encryption and multi-factor authentication. The standard also expects you to detect and disable active vendor sessions. Plan the support path for the SIEM itself, including who can open a session, how it is logged and how it is ended. Confirm with your compliance team which assets are in scope before designing this path.
Scope by program and sector
NIST SP 800-82 Rev. 3 is the NIST guide to operational technology security and is a useful reference for OT networks whatever the regulator. Whether CMMC under 32 CFR Part 170 or FedRAMP applies depends on the program and contracts you work under, and FedRAMP's scope guidance should be read with your authorizing body. Confirm these obligations with your compliance lead before selecting controls, because they set the evidence you must keep.
Which open-source platforms fit an offline SIEM
This table compares Wazuh, OpenSearch, Graylog Open and Security Onion on offline operation, licensing and gaps for isolated networks. Read each row against your own classification, OT and audit requirements, and confirm the items marked for verification with each vendor before you commit to a design.
| Platform | Offline operation | Trade-offs and what to confirm |
|---|---|---|
| Wazuh | Publishes an offline installation guide for systems without internet connectivity and can use an offline vulnerability repository in place of the Wazuh CTI platform | Four components (indexer, server, dashboard, agents) to stage and patch offline; confirm agent OS coverage and how the offline vulnerability feed is refreshed |
| OpenSearch | Storage and search layer under Wazuh; runs on hosts you provision inside the isolated network with no vendor cloud dependency | Cluster sizing, index lifecycle and backups are your responsibility; confirm the Wazuh version pairing with the OpenSearch build you stage |
| Graylog Open | Free to self-manage under the Server Side Public License with no ingestion caps; you control the installation media and hosts | Correlation and threat detection depend on Graylog Security, a paid tier; confirm which SIEM features your use case needs and whether Graylog Enterprise is required for offline support |
| Security Onion | Full installs without internet access; the ISO carries packages plus NIDS, YARA and Sigma rules, and soup updates from newer ISO media in airgap mode | Repository is under the Elastic License 2.0, not an OSI open-source license; OT protocol coverage is a gap to verify against your asset list; Security Onion Pro is a paid tier |
How to deploy Wazuh fully offline
This sequence produces a Wazuh deployment that runs, detects and reports inside an isolated network, with every package, rule set and vulnerability dataset checked and carried across the boundary by a documented process.
Build offline media on a staging network
On a connected staging network, your team downloads the Wazuh packages, the indexer, server and dashboard components, the agent packages for each endpoint type, and any operating system dependencies the target hosts will need. Record the version, checksum and source of every file in a manifest kept under change control. The step is complete when the manifest matches the media set and an independent reviewer has checked the hashes before anything is staged for transfer.
Follow the Wazuh offline installation guide
Using the Wazuh offline installation guide, the platform engineering team installs the indexer, server and dashboard on hosts inside the isolated network from the staged media. Confirm the installation order, certificate generation and service start-up steps against the guide for the exact release you are deploying, since the procedure is version-specific. The step is complete when all three components report healthy services and the dashboard loads from an administrator workstation inside the enclave.
Load the offline vulnerability repository
Vulnerability detection correlates agent software inventories with vulnerability data. In a disconnected network the Wazuh Cyber Threat Intelligence platform is not reachable, so the team loads an offline vulnerability repository built on the staging network and configures the vulnerability detection module to use it. Confirm with Wazuh which data sources the offline repository covers and how it is refreshed for your environment. The step is complete when a test endpoint's software inventory produces vulnerability findings from the local data.
Mirror container and Kubernetes images
If any Wazuh component or supporting service runs as a container or on Kubernetes, the team pulls every required image on the staging network and pushes it to a private registry inside the enclave. For Kubernetes control planes, kubeadm can list and pull the required images in advance, and the image repository flag points the cluster at the internal registry. The step is complete when every image referenced by the deployment manifests resolves from the private registry with no outbound network access.
Install agents from verified packages
Endpoint administrators install Wazuh agents from the agent packages carried across the boundary, not from any internet source. Before deployment, the security team verifies each package against the checksums in the manifest and records the verification result, which supports supply chain integrity evidence of the kind NERC CIP-013-2 asks utilities to address. Agents are enrolled against the internal server. The step is complete when each enrolled agent appears in the dashboard with its expected operating system and version.
Set up a controlled transfer for updates
Establish a written cross-domain cycle that moves updates, rules and vulnerability data from the staging network into the enclave through an approved transfer method. Each cycle repeats the same steps: download, verify, scan, transfer, install and record. The Wazuh release notes list patch releases for the 4.14 line at intervals of roughly one to two months, so the cycle should be repeatable on that cadence. Confirm with your accreditation or compliance owner how the transfer is authorized. The step is complete when a dry run of the full cycle is documented and signed off.
How the design changes for classified, OT and disconnected sites
Classified enclaves usually restrict what can enter and leave the network, so the first design question is how the Wazuh components, agent packages and vulnerability data reach the enclave at all. Plan the deployment around the offline installation path rather than assuming a pull from public repositories. Agents on Linux, Windows, macOS, Solaris, AIX and HP-UX can all report to a server inside the enclave, which keeps endpoint coverage inside the boundary.
- Agent coverage: Confirm which operating systems and versions in the enclave are on the supported agent list, and which legacy hosts need agentless collection instead.
- Offline packages: Stage every package, rule set and vulnerability repository through the approved transfer process, and keep a record of what was imported and when.
- Vulnerability data: Use an offline vulnerability repository in place of the Wazuh Cyber Threat Intelligence platform, and document its update source and refresh process.
- Accreditation scope: Ask the security authorizing official whether the SIEM, its dashboard and its indexer are inside the system boundary, and how logs from other enclaves are handled.
- Software integrity: Verify checksums or signatures of each imported component before installation.
Confirm with the enclave's authorizing official and security office which transfer method, media handling rules and accreditation evidence apply before any deployment starts.
OT plant networks run controllers and engineering stations that often cannot host third-party agents, and the vendors may not support them. Many sites therefore rely on agentless monitoring, where firewalls, switches and controllers send syslog to a collector inside the plant zone, and the Wazuh server periodically probes APIs where they exist. Where a host is supported, a Wazuh agent can run on it, but agent installation on production controllers needs the asset owner's approval and a change window.
- Agent coverage: List the plant hosts that run supported operating systems, such as Windows engineering workstations, and mark each as agent-capable or not.
- Agentless syslog collection: Plan syslog from firewalls and controllers to a collector in the plant zone, with defined ports, message formats and time synchronization.
- Periodic API probing: Confirm which devices expose APIs the Wazuh server can query, and what credentials and read-only accounts those calls require.
- Passive monitoring: Ask the OT team whether any active scanning or polling is allowed, since probes can disrupt fragile devices.
- Zone boundaries: Confirm how the monitoring data crosses from the plant zone into the enterprise zone.
Confirm with the asset owner and plant operations team which devices may be contacted, at what rate, and during which maintenance windows, and align the design with NIST SP 800-82 Rev. 3, the OT security guide.
Disconnected remote sites, such as mines, pipelines, field offices or vessels, may have no link for days and only a periodic transfer window. The design must tolerate delayed delivery, store events locally and carry updates on physical or scheduled media. A local Wazuh server with the indexer and dashboard can keep alerts and compliance views on site, while summaries or exported alerts move to the central team during transfer windows.
- Local buffering: Confirm the agents and server hold events long enough to cover the longest expected gap, and that disk sizing accounts for that backlog.
- Transfer media: Define what moves on each transfer, how it is signed and who verifies it on arrival.
- Offline updates: Plan rule, decoder and vulnerability data updates as offline bundles, with a fixed cadence agreed with the central team.
- Time accuracy: Verify time sources on site, because timestamps drive correlation when logs arrive in batches.
- Local response: Decide which alerts must be actioned on site without waiting for the central team.
Confirm with the site operator and central security team how long the site must operate without contact, what must be reported after reconnection, and who is accountable for the transfer media.
Evidence auditors and regulators ask for
An isolated Wazuh deployment is only as credible as the records behind it. The cards below cover the artifacts that show what crossed the boundary, what the platform saw, and who could reach it, so that an assessor can trace each claim to a record.
Offline media manifest and hash records
Every package, rule set and vulnerability data file carried across the boundary should appear in a manifest with its file name, version, source and cryptographic hash, recorded before the media is opened on the isolated side. Verify the hashes again after transfer and keep both sets of values. This is the evidence that supports NERC CIP-013-2 software integrity and authenticity checks. Ask your integrator how the manifest is produced and who signs it.
Change records for each transfer
For each media movement, keep a change record that names the requester, approver, transfer date, media identifier, scan results and the person who imported the content. Illustrative example: a Wazuh server package and a rules update moved on one approved USB device, with the manifest attached. Auditors use these records to test whether transfers follow your own change process rather than ad hoc copying.
Offline vulnerability repository update history
Wazuh vulnerability detection can use an offline repository in place of the Wazuh CTI platform, so the update history of that repository becomes the source record for every vulnerability finding. Log each import with its date, source snapshot and hash. Confirm with your integrator how stale data is flagged, because a finding is only as current as the last import.
Agent inventory and configuration baselines
Keep a current inventory of every Wazuh agent with its operating system, version, group membership and assigned configuration. Store a baseline of each configuration and record every deviation. Agents on Linux, Windows, macOS, Solaris, AIX and HP-UX should all appear, including OT hosts that must be recorded even when they are monitored only through syslog. The inventory is also the input to vulnerability and compliance reports.
Remote access logs and session controls
Remote access into an isolated enclosure is the control most often examined. Record every session with the user, time, source, purpose and the approval that allowed it, and show that access passes through an intermediate system with encryption and multi-factor authentication. NERC CIP-005-7 sets these expectations for high and medium impact systems, and you should also confirm how active vendor sessions are detected and disabled.
Compliance dashboard outputs
The Wazuh dashboard includes prebuilt views for PCI DSS, GDPR, HIPAA and NIST 800-53. Export and archive these outputs on a schedule, with the rule set version and data date stated, so that a point-in-time view can be reproduced later. Treat dashboard output as supporting evidence only. Your assessor will decide which controls it satisfies and whether the mapping fits your scope.
Questions buyers ask before an offline SIEM project
What does an offline Wazuh project cost and what drives the price?
The cost depends on the written scope, not on a fixed package. The main components are the engineering and implementation effort, the hardware or virtual machines for the Wazuh indexer, server and dashboard, storage sized for your retention period, the agent rollout across endpoints, and the internal staff time your team spends on access, change approval and testing. Software licensing is not a line item because Wazuh is free and open source under GPLv2. Ask for a quote that separates one-time work such as design, installation, offline packaging and integration from recurring work such as upgrades, patch cycles and support. DeployOpen does not publish a price list. A quote follows from the written scope, which should state the number of endpoints, log sources, retention and the transfer process into the isolated network.
How many endpoints and log sources change the size of the project?
Endpoint count and log volume are the two largest sizing inputs. Agents run on Linux, Windows, macOS, Solaris, AIX and HP-UX, and agentless devices such as firewalls can send syslog. Each platform type adds packaging, testing and configuration work, and each additional source adds decoders or rules to review. The Wazuh server can run as a cluster for horizontal scaling, so larger estates may need more nodes. Ask for a sizing review that records events per second, retention period, alert volume and the number of distinct operating systems before any hardware is ordered.
Can our own team install and run Wazuh offline without outside help?
A capable internal team can do it, but the work is more than installing packages. Teams usually need people who understand Linux administration, certificates, indexer tuning, agent deployment and change control, plus someone who owns the offline update process. DeployOpen plans, implements, hardens, backs up, upgrades and documents self-hosted platforms in your own infrastructure, including isolated networks, and can work as a fixed-scope engagement or a retainer. Decide early which tasks stay in-house, such as day-to-day alert triage, and which need a partner, such as the first build, the transfer procedure and the first offline update cycle.
How long does an offline Wazuh deployment take?
The duration depends on scope, the number of environments and how quickly approvals and transfer media can move through your accreditation process. No general timeline is supplied here, so treat any single number with caution. Phases that usually take the most time are the network and access review, building the offline media set, validating each package before it crosses the boundary, and agent rollout in change windows. Ask for a written plan with milestones, dependencies on your security and change boards, and the testing you expect before production sign-off.
What do we need from our staff during an air-gapped project?
Your staff need to provide access, approvals and operational knowledge. Expect requests for network diagrams, asset inventories, the list of log sources, retention requirements, identity and certificate standards, and named owners for change control. Security and compliance staff should confirm how software and patches are verified before entering the isolated network, which aligns with the supply chain control in NERC CIP-013-2 for utilities. Operations staff will be needed for agent deployment windows and for testing alerts against real activity. Agree these responsibilities in the written scope so that delays are visible.
Does Wazuh need internet access after the installation is complete?
Wazuh's offline installation guide covers systems without internet connectivity, so a deployment can be built and run without a live connection. The point to confirm is which features depend on outside services. Vulnerability detection normally correlates agent software inventory with data from the Wazuh Cyber Threat Intelligence platform, but it can use an offline vulnerability repository instead. Any feature that calls an external update or lookup service must be replaced or disabled in your design. Review the offline installation guide and the vulnerability detection documentation with your engineers before committing to an architecture.
Is Wazuh free to use under the GPLv2 license, and what does the OpenSSL exception mean?
Wazuh is free and open source and is distributed under the GNU General Public License version 2. The license includes an exception that permits linking with the OpenSSL library. In practice this means you can deploy and modify Wazuh without a per-GB or per-endpoint license fee, but you still need to meet GPLv2 obligations if you distribute modified copies. Have your legal or procurement team read the LICENSE file in the Wazuh repository and confirm how it applies to your internal use, any packaged builds you create for the isolated network, and any third-party components you add.
What does the offline vulnerability repository not cover?
An offline vulnerability repository replaces the online Wazuh Cyber Threat Intelligence feed as the source of vulnerability data, so it only knows what it was given at the time of the last transfer. It will not reflect newer advisories until a new data set is imported, which means there is a lag between vendor disclosure and detection. It also depends on the accuracy of the software inventory that agents report, so unmanaged or unusual software may be missed. Confirm with your team how often the repository will be refreshed, who checks each import for integrity, and which assets sit outside agent coverage, since those need separate scanning.
What changes for teams moving off Sentinel
The data work starts with an inventory of what Sentinel collects today. Each Azure data connector, whether it pulls from Microsoft services, syslog or an API, needs a named replacement on the Wazuh side: a Wazuh agent on Linux, Windows or other supported endpoints, or syslog collection and API probing for devices such as firewalls that cannot run an agent. Confirm which sources can carry an agent inside the air-gapped boundary, because OT and classified hosts often cannot accept new software without a change approval. Analytics rules need the same treatment. KQL detections do not transfer as written, so each one must be rebuilt as a Wazuh rule with its decoders and tested against replayed events before cutover. Export the rule definitions from Sentinel before the tenant is retired, and record which rules you are keeping, rewriting or dropping. The sibling page on migrating from Microsoft Sentinel to Wazuh covers the export jobs and rule translation in detail.
Historical data is the constraint specific to isolated sites. An online migration can stream logs out of Azure while the new platform is being built, but a disconnected network may have no path from Azure to the target at all. Decide early how history will cross the boundary: a one-way transfer on approved media, a data diode, or a decision to keep the old exports as read-only evidence outside the Wazuh indexer. Each option has its own review and sign-off, and every transferred file should pass the integrity and provenance checks your change process requires. Ask whether your accreditor or asset owner needs retention of the original Sentinel records or only the alerts and cases. Plan the transfer before the cutover date, not after it.
How DeployOpen deploys open-source SIEM in isolated networks
DeployOpen plans, implements, hardens, documents and supports self-hosted open-source platforms inside the client's own infrastructure, including isolated networks. For an air-gapped Wazuh deployment, the work starts with the boundary, the package transfer path and the vulnerability data the site will use, then moves to a tested offline build that the client's own team can operate.
How we scope
- Written scope agreed before work starts, listing hosts, enclaves, log sources and retention for the isolated network
- Mapping of each data source to a Wazuh agent, syslog path or agentless collection method
- Decision on offline vulnerability repository content and how updates will cross the boundary
How we build
- Offline installation of the Wazuh indexer, server and dashboard, following Wazuh's offline installation guide
- Verification of every package, image and rule set before it is carried across the isolated boundary
- Testing of agent enrollment, alert flow and compliance dashboards inside the disconnected environment
How we hand over
- Deliverables documenting architecture, offline update procedure, backup and restore steps and access controls
- Evidence packs organized for the client's assessors, prepared for independent auditors to review
- Knowledge transfer so the client's operators can run upgrades and troubleshoot without outside access
How we support
- Ongoing support, upgrades and continuous testing available as a retainer after handover
- Periodic offline update cycles planned around Wazuh patch releases and the client's change process
- Confirmation of support scope, response expectations and retainer terms in writing with the client

