Red Hat releases OpenShift 5 with AI troubleshooting that leaves people in control

Red Hat is giving AI a larger role in investigating infrastructure problems, but not the final say on fixing them. OpenShift 5, released on November 9 at KubeCon + CloudNativeCon North America, adds agent-based troubleshooting and AI-assisted upgrade planning. The generally available release keeps administrators in control of whether recommended corrective actions are executed.

This is the platform’s first major version upgrade in seven years. OpenShift is Red Hat’s Kubernetes-based application platform, bringing containers, virtual machines and AI workloads into a common environment. Version 5 combines changes to day-to-day operations with new security controls and a different path for moving existing installations to the new release.

Investigating a problem, not automatically fixing it

The operational changes build on OpenShift Lightspeed, the platform’s generative AI assistant. Its new troubleshooting capability investigates events in a cluster, examines related information and works through possible causes before proposing remediation steps. That makes the agent an investigator and adviser; the administrator decides whether to carry out the proposed action.

Lightspeed also lets administrators request performance visualizations in ordinary language. An administrator can ask to see CPU utilization over the previous week, add memory consumption and overlay application errors. Those charts can then be saved to dashboards that update continuously, turning a one-time request into an ongoing view of system behavior.

Checking what an upgrade might break

The AI-driven upgrade planner checks for compatibility problems before an update begins. It examines three parts of the environment: application dependencies, Kubernetes application programming interfaces and optional operators.

Separately, Red Hat says OpenShift 4 customers can install version 5 as an in-place update, without reinstalling software or migrating workloads. The two changes address different parts of upgrading: the planner checks the software environment, while the in-place path changes how customers move to the new major version.

The release also introduces a revised lifecycle approach that Red Hat describes as providing greater predictability for customers running long-lived enterprise applications and virtualized infrastructure.

Security boundaries and shared computing resources

OpenShift 5’s security changes span separation between customers, network controls and the platform’s own communications. They include hardened multitenancy—separation between tenants sharing the platform—expanded network isolation, automated certificate management and post-quantum cryptography for control-plane communications.

Red Hat is aiming those controls partly at sovereign cloud providers and regulated organizations that need strict tenant separation and evidence of compliance.

The platform also supports automated GPU resource allocation across AI inference—the work of running models—and traditional enterprise workloads. That capability is intended to improve use of expensive computing resources; it is not a reported measurement of savings or utilization gains from this release.