Air-gapped monitoring
Metrics, logs, traces and alerts for networks that cannot reach the internet. We design, install and run the whole stack offline: on-prem, in a no-egress cloud network, or in a sovereign or disconnected cloud.
We reply within one business day.
Zero egress
no telemetry, update checks or licence calls leave the network
From scratch
design, offline install, handover and run
Middle East
air-gapped monitoring projects delivered
Connected staging → controlled transfer → offline registry and repositories → monitoring stack inside the air gap
Connected staging → controlled transfer → offline registry and repositories → monitoring stack inside the air gap
Works with
Prometheus
Grafana
Zabbix
Elasticsearch
OpenSearch
Kibana
ClickHouse
OpenTelemetry
Kubernetes
Rancher / RKE2
Ansible
Vault
VMware
Proxmox
In short
We build air-gapped monitoring and logging platforms from scratch with open-source tools: Prometheus or Zabbix for metrics, Loki or Elasticsearch for logs, Tempo for traces, Grafana for dashboards and alerts, and OpenTelemetry to collect data. Everything installs and updates without internet access, through mirrored package repositories, an internal container registry and signed artefacts brought in through your approved transfer process. Nothing phones home. We have delivered air-gapped monitoring projects in the Middle East and work on-prem, in isolated cloud networks and in sovereign clouds.
Services
Metrics, logs, traces and alerting designed for your scale, retention and network zones, with high availability where it matters.
Private mirrors for everything the stack needs: OS packages (apt, yum/dnf), container images, Helm charts, Python and npm packages, Grafana plugins and Elastic artifacts, in Harbor, Nexus, Artifactory or Pulp.
Artefacts scanned, signed and listed outside the gap, brought in through approved media or a data diode, and verified inside.
Prometheus or Mimir, Zabbix for network devices and classic servers, Loki or Elasticsearch for logs, Tempo for traces, Grafana on top.
Usage reporting, update checks, plugin catalogues and online licence checks disabled and verified. Offline licences where a product needs one.
Ansible, Terraform and Helm so the whole stack can be rebuilt, audited and copied to another site.
Alerts to on-prem email, chat or SMS gateways, with owners and runbooks, and no dependence on cloud paging services.
A tested, repeatable process to bring in security updates and new versions on a schedule your security team signs off.
Runbooks and training for your team, or ongoing operation by our engineers under your access rules.
Examples
Monitoring and logging for a new secure environment, designed and installed with no internet access at any point.
Central logs from servers, network devices and applications, kept for the period your regulator requires.
Cluster and workload monitoring for Kubernetes that pulls everything from an internal registry.
The monitoring we build becomes the foundation for AI agents that handle routine SOC and NOC work, governed through Opsfinitive.
Remote sites monitored through local proxies, with a central view where the network design allows it.
How to start
Requirements and constraints
Target architecture and sizing
Transfer process design
Plan and timeline
Offline repositories and registry
Monitoring and logging stack
Dashboards and alerts
Runbooks and handover
Patch and upgrade cycles
Capacity and alert tuning
New sources and dashboards
Response times in writing
Process
What must be monitored, retention, network zones, compliance rules and how data may enter the network.
Mirrors and registry first, then the stack, installed and tested inside the air gap.
Runbooks, training and a tested update process, or ongoing operation by us.
Why it is different
Most monitoring tools assume they can download updates, plugins, licences and container images, and send usage data home. Inside an air gap, every one of those assumptions breaks.
Packages, container images, Helm charts, Grafana plugins and dashboards all need a mirrored, offline source.
Security patches have to come in on a schedule, through a process your security team approves and can audit.
Usage reporting, update checks, plugin catalogues and online licence checks must be found and switched off.
Time sync, DNS, certificates, identity, object storage and backups all run inside the network.
Who brought what in, when, and who approved it. The transfer process itself has to be logged.
Air-gapped builds fail on details. We have delivered them and plan for those details from day one.
An honest note before you book.
Data residency is the only requirement: a monitoring stack in your own region with outbound access may be enough.
A small environment where a managed, in-country SaaS meets your rules.
You need only a few servers monitored for a short period.
Where it runs
A no-egress cloud network is often called air-gapped, but strictly it is network-isolated: the provider still runs the control plane. Many regulators accept it; some require a full air gap. We help you decide which one your rules require.
FAQ
Straight answers. Ask us anything else on a call.
Monitoring, logging and alerting that run entirely inside a network with no connection to the internet. All software, updates and dashboards are brought in through a controlled process, and no data leaves the network.
Partly. In a public cloud you can build a network with no internet gateway, no outbound access and private endpoints only. Strictly that is network-isolated rather than air-gapped, because the provider still runs the control plane, but many regulators accept it. A full air gap needs on-prem hardware or a disconnected or sovereign cloud offering. We build the same monitoring stack in all three.
Prometheus, Grafana, Loki, Tempo, Zabbix, OpenSearch, ClickHouse, Fluent Bit and the OpenTelemetry Collector all run without internet access. Elasticsearch runs offline too: the free tier needs nothing, paid features use a licence file uploaded offline, and Fleet integrations come from a self-hosted Elastic Package Registry. We disable usage reporting, update checks and plugin downloads in every component.
Elasticsearch or OpenSearch when people need full-text search across logs and security analytics (an EFK stack). OpenSearch is fully Apache 2.0 with no licence file. Loki when logs are mainly read next to metrics in Grafana and storage cost matters. ClickHouse when volumes are very large and you query logs like analytics data. Our Elasticsearch specialists at elasticsearch.expert cover search, OpenSearch and ClickHouse in depth.
Yes. Fleet Server runs inside the network; integrations come from a self-hosted Elastic Package Registry and agent binaries from a self-hosted Elastic Artifact Registry, both refreshed through the same controlled transfer as everything else. Kibana is pointed at these internal sources, so agents enrol and upgrade without internet access.
Through a transfer process: artefacts are downloaded, scanned and signed outside the air gap, moved in on approved media or through a data diode, verified, and loaded into internal mirrors. Each transfer is logged.
Yes. We design and build the whole platform, from mirrors and registry to dashboards and alerts, and either hand it over with runbooks or run it for you.
Not necessarily. We can work on site, through your approved remote access, or prepare everything outside the air gap for your team to bring in and install with us guiding.
Yes, if they are governed. In air-gapped projects the same question keeps coming up: once every alert and log is in one place, can AI handle the routine work of the night shift without anything leaving the network? That is the question Opsfinitive was built to answer. Founded by our co-founder Davis Mathew, its Plumbline layer runs AI agents inside sovereign and air-gapped environments, controls what each agent may do and records every action for audit. It is vendor-neutral and delivered in the GCC and EU.
We have delivered air-gapped monitoring projects in the Middle East for organisations that could not use cloud monitoring. Client names stay confidential.
Next step
Send a few details, and pick up to 3 times if you would like a call. We reply within one business day.
You talk to the expert who would do the work
Written scope and price after the call
No obligation