| name | ocp-lightspeed-configure |
| metadata | {"author":"rhoai3-coding-demo","version":"1.0.0","platform-family":"ocp","platform-baseline":"repo","ocp-baseline":"repo","skill-group":"OpenShift Platform"} |
| description | Use when configuring Red Hat OpenShift Lightspeed 1.0, including OLSConfig customization, LLM provider settings, conversation history, telemetry, TLS certificates, RBAC, and user access controls. Do NOT use for concepts (use ocp-lightspeed-about), installing (use ocp-lightspeed-install), operations (use ocp-lightspeed-operate), or troubleshooting (use ocp-lightspeed-troubleshoot).
|
OCP Lightspeed Configure
Use this skill to configure Red Hat OpenShift Lightspeed 1.0 grounded in
official product documentation.
Source Grounding
Read references/source-capture.md before using product behavior. Official Red
Hat documentation is product authority. This skill covers OLSConfig CR
customization, LLM provider credential and endpoint setup, TLS configuration,
RBAC, query filtering, RAG databases, cluster interaction (MCP), token quotas,
and PostgreSQL persistence.
Key Configuration Areas
- Credentials secrets — API token secrets per LLM provider in
openshift-lightspeed namespace (OpenAI, RHOAI, RHELAI, Azure, watsonx,
Google Vertex AI).
- OLSConfig CR — Single cluster-scoped resource (
ols.openshift.io/v1alpha1,
name: cluster) that drives the entire service deployment.
- LLM provider configuration — Provider type, URL, models, and
credential references under
spec.llm.providers[].
- TLS certificates — Custom TLS for the backend endpoint
(
spec.ols.tlsConfig) and trusted CA for LLM providers
(spec.ols.additionalCAConfigMapRef).
- RBAC and user access — ClusterRole
lightspeed-operator-query-access
bound to users or groups.
- Query filters — Regex-based sensitive data redaction before LLM
submission (
spec.ols.queryFilters).
- BYO Knowledge (RAG) — Custom RAG databases via container images
(
spec.ols.rag).
- Cluster interaction (MCP) — Observability MCP server for cluster
introspection (
spec.ols.introspectionEnabled), custom MCP servers
(spec.mcpServers, feature gate MCPServer).
- Token quotas — Per-user and per-cluster token limits
(
spec.ols.quotaHandlersConfig).
- PostgreSQL persistence — Conversation history storage
(
spec.ols.storage).
- Google Vertex AI —
google_vertex and google_vertex_anthropic
provider types with GCP service account authentication.
- Tool filtering — Query-based tool selection for MCP
(
spec.olsConfig.toolFilteringConfig, feature gate ToolFiltering).
OLSConfig CR Summary
- apiVersion:
ols.openshift.io/v1alpha1
- kind:
OLSConfig
- metadata.name:
cluster (singleton, cluster-scoped)
- Required spec fields:
spec.llm and spec.ols
- Key spec.ols fields:
defaultModel, defaultProvider,
additionalCAConfigMapRef, byokRAGOnly, conversationCache,
deployment, introspectionEnabled, logLevel, proxyConfig,
queryFilters, quotaHandlersConfig, rag, storage, tlsConfig,
tlsSecurityProfile, userDataCollection
See references/official-doc-extraction.md for complete field reference and
YAML examples.
LLM Provider Configuration
Supported provider types:
openai — OpenAI API
azure_openai — Microsoft Azure OpenAI (supports Entra ID auth)
watsonx — IBM watsonx (requires projectID)
rhelai_vllm — Red Hat Enterprise Linux AI vLLM
rhoai_vllm — Red Hat OpenShift AI vLLM
google_vertex — Google Vertex AI (Gemini)
google_vertex_anthropic — Anthropic on Google Vertex AI
All providers require a credentials Secret in namespace openshift-lightspeed.
The secret key is always apitoken (except Google Vertex AI which uses a GCP
service account JSON key via credentialKey).
RBAC and Access Control
OpenShift Lightspeed RBAC is binary — the Operator creates a ClusterRole
lightspeed-operator-query-access. Grant access via:
oc adm policy add-cluster-role-to-user lightspeed-operator-query-access <user>
oc adm policy add-cluster-role-to-group lightspeed-operator-query-access <group>
The kubeadmin account always has access. Cluster introspection uses MCP
server authentication (file-based secrets or Kubernetes passthrough).
Workflow
- Read
references/official-doc-extraction.md for exact YAML patterns.
- Identify the configuration task:
- Initial OLSConfig creation for a specific LLM provider
- Adding/changing TLS certificates
- Granting user or group access (RBAC)
- Configuring query filters for data redaction
- Enabling BYO Knowledge RAG databases
- Enabling/disabling cluster interaction (MCP)
- Configuring token quotas
- Enabling PostgreSQL persistence
- Exposing the service via a Route
- Configuring Google Vertex AI
- Use the exact
apiVersion, field names, and namespace from the extraction.
- Validate with the verification commands documented per section.
- Never invent CR fields not documented in the OLSConfig API reference.
Related Skills
ocp-lightspeed-about — OpenShift Lightspeed concepts and architecture
ocp-lightspeed-install — Operator installation procedures
ocp-lightspeed-operate — Day-2 operations and maintenance
ocp-lightspeed-troubleshoot — Diagnostics and recovery
ocp-lightspeed-upgrade — Upgrade procedures
ocp-lightspeed-uninstall — Removal procedures
References
references/source-capture.md
references/official-doc-extraction.md