
Migration
Migrate NSX and firewall policy to Isovalent and OpenShift.
One router replaces three sync jobs. It reads NSX DFW and third-party firewall policy (Palo Alto, Check Point, Fortinet, Cisco ASA/FTD) into Fabrix, converts it to label-native Cilium network policy, and publishes it, with simulation before every run.

How a migration runs
Six steps, from baseline to sign-off.
As-Is Report
Document the source estate before anything changes: sections, rules, groups, members and objects, exported as the migration baseline.
Endpoint & Label Sync
VMs not yet migrated become label-selectable endpoints on Isovalent and OpenShift; NSX tags are stamped as labels on migrated pods and VMs.
CIDR Sync
Raw-IP and IP-set groups become CiliumCIDRGroups. Bidirectional mode writes child groups back to NSX during coexistence.
Policy & Cleanup
Rules convert to CNPs with ports and direction preserved. Cleanup removes temporary CIDR groups and endpoints once peers are label-covered.
Validate with Hubble
Fabrix's Hubble integration streams live flows onto the topology by verdict, so dropped traffic is caught and fixed in one click.
As-Built Report
Document the target as published: CNPs, CIDR groups, endpoints and per-rule coverage, ready for sign-off against the as-is baseline.
- Deny AnalyzerPreserves NSX first-match order under Cilium's absolute denies: redundant denies are skipped, true carve-outs pushed, order conflicts flagged for review.
- Simulate firstEvery rule is classified Safe, Warn or Block by the same engine the live run uses. Risky rules are skipped on publish.
- Safe to runTouches only Fabrix-managed objects, continues past per-rule failures, audits every write, and splits logs and diagnostics per sync type.
Hubble flow validationConfirm target cluster health
Everything Fabrix does
Choose a role in "Your role" (top right) and this page reshapes around what you care about.
Migration tools in the console




Also for migration teams
Around the move.
- Policy-Gap RemediationA dropped flow prefills an allow; an unexpected allowed flow prefills a deny.
- KubeVirt VM SupportVMs moved to OpenShift Virtualization are managed like pods, in inventory and topology.
- Applied Policy InspectorSee the eBPF policy actually enforced on any endpoint after publish.
- Prism ApplianceRehearse a migration at target scale on a single node before touching production.
- Workload ConsoleExec into pods or open a VM serial console from the browser, RBAC-gated and audited.
- Cluster MeshSee cross-cluster links where policy references workloads in more than one cluster.