Self-managed Lightspeed moves OpenShift health analytics inside the air gap

Red Hat has published a Developer Preview of Red Hat Lightspeed advisor self-managed, an Advanced Cluster Management add-on that runs health recommendations and upgrade-risk analysis without sending cluster data outside a disconnected environment. The preview matters less as a new dashboard than as an architectural change: it relocates the advisor pipeline from Red Hat’s hosted service onto the customer’s management hub.

What changed

In the connected service, the Insights Operator collects cluster-state archives and sends them to the Hybrid Cloud Console for processing. That path does not work for clusters isolated by network policy, regulation or data-residency requirements.

The self-managed preview redirects the Insights Operator on each managed cluster to a local endpoint on the Red Hat Advanced Cluster Management hub. The hub processes the archive, stores the results and exposes recommendations through the existing Advanced Cluster Management console. Upgrade-risk predictions also use metrics already held locally in the hub’s Thanos observability service.

Red Hat says the deployed system makes no external API callbacks and sends no telemetry upstream. That gives platform teams a concrete separation between operational analysis and outbound connectivity rather than asking them to treat obfuscation as a substitute for locality.

Why the deployment model is different

Red Hat’s hosted advisor uses horizontally scalable microservices for a multitenant service. The self-managed edition instead packages the processing path as a single deployment serving one Advanced Cluster Management hub and its spoke clusters.

That trade-off is deliberate. Consolidating the functions into one pod reduces idle resource overhead and removes queues, serialization and network calls between services. For a disconnected fleet, where the scale boundary is known and the hub already concentrates management data, a compact local pipeline is a more practical fit than reproducing the hosted architecture.

The design also makes Advanced Cluster Management the control point. Cluster archives arrive through the hub, observability data is already present in Thanos, and operators read the results in the same console they use for fleet management. Teams do not have to introduce a second interface or a separate cloud exception simply to obtain advisor findings.

Who should evaluate it

The immediate audience is platform teams operating OpenShift in air-gapped or tightly regulated environments that already use Advanced Cluster Management. They can test whether local recommendations identify configuration and update risks without weakening their network boundary.

The qualification is important: Red Hat labels this release Developer Preview. It is not yet an Advanced Cluster Management Technology Preview or generally available feature, so it should be evaluated as design feedback and pre-production validation, not treated as a supported production control.

Teams evaluating the preview should verify the resource footprint on the hub, archive flow from spoke clusters, data-retention behavior and the usefulness of the resulting recommendations against their existing change process. The larger signal is clear: Red Hat is testing a way to make its proactive OpenShift health analysis deployable where a hosted dependency is unacceptable.