Red Hat OpenShift Lightspeed 2 adds live and autonomous troubleshooting

Red Hat OpenShift Lightspeed 2 can now investigate OpenShift cluster problems directly, either through a live troubleshooting chat or an autonomous agentic run. The update gives the system read-write tools, while requiring explicit user approval before it changes a cluster.
- OpenShift Lightspeed 2 offers live troubleshooting chat and autonomous agentic runs.
- Troubleshooting chat is generally available.
- Autonomous agentic runs are a Technology Preview.
- Cluster changes require explicit user approval.
- The update adds read-write tools to OpenShift Lightspeed.
From documentation answers to cluster investigations
OpenShift Lightspeed began as a chat assistant inside the Red Hat OpenShift console. Version 1.0 answered operational questions by searching Red Hat documentation and matching responses to the OpenShift version running on a cluster. It could explain how to expose an application with a route or provide guidance for a pod stuck in a crash loop, but users had to carry out the investigation themselves.
Version 1.1 added a built-in Model Context Protocol server, giving OpenShift Lightspeed direct read access to cluster data including pods, events, logs and resources. It could inspect events and logs and report what it found, while users still had to make any changes. Conversations could also include items such as a Prometheus alert.
Red Hat says later releases added Google Vertex AI and Amazon Bedrock as supported model providers, automatic detection for Red Hat OpenShift Service on AWS clusters, OpenTelemetry tracing and configurable TLS profiles. The company presents those additions as the foundation for the operational capabilities in version 2.
Two ways to troubleshoot in OpenShift Lightspeed 2
The troubleshooting chat is generally available in OpenShift Lightspeed 2. It works with the user during a live investigation, listing pods, reading events, checking resource limits and showing each step as it happens. Users can redirect the investigation, such as asking it to check another namespace or examine node conditions.
Actions that would change the cluster pause for approval. OpenShift Lightspeed 2 can also visualize selected cluster metrics as dynamic graphs in the conversation through its Perses Dashboards integration.
The second option, autonomous agentic runs, is available as a Technology Preview. The user points OpenShift Lightspeed at a problem and asks for an autonomous run. The system creates an isolated execution environment on the cluster, with its own workspace and OpenShift command line access. It then builds a plan, checks the cluster step by step and adjusts its approach until it reaches an answer. The result includes a root cause and a proposed fix, which the user decides whether to approve.
What the update means for OpenShift operators
OpenShift Lightspeed 2 moves the product from explaining how to implement a fix toward implementing fixes itself under human control. That could reduce the amount of manual checking required during cluster incidents, while preserving a user decision before changes are made.
Red Hat says it is considering future capabilities including memory across autonomous runs, access through the command line, chat tools and an API, integration with other tools and agents, and support for procedures encoded by an SRE team. The company is also exploring investigations across managed clusters with Red Hat Advanced Cluster Management for Kubernetes. Red Hat describes these as future opportunities, rather than shipped capabilities.
Red Hat: What's new in Red Hat OpenShift Lightspeed 2


