Practice for the Red Hat Certified Specialist in OpenShift AI (EX267) exam with 17 exam-style practice questions, instant answer reveals, and concise explanations of every correct answer. Topics include: WHAT COMPONENT DEPLOYS MULTI-TENANT ROUTING ARCHITECTURES AND DECLARES INGRESS POLICIES SECURING THE OPENSHIFT AI KSERVE. Follow @CertPunch and visit certpunch.com for more certification practice exams and study content.
Prefer hands-on? Take this round as an interactive practice test — answer every question, get instant feedback, and see your score: Start the Red Hat Certified Specialist in OpenShift AI (EX267) practice test →
What you will practice
- WHAT COMPONENT DEPLOYS MULTI-TENANT ROUTING ARCHITECTURES AND DECLARES INGRESS POLICIES SECURING THE OPENSHIF…
- Which tool tracks a comprehensive dependency inventory to ensure pre-built Jupyter containers comply with sys…
- WHAT FIELD ASSIGNMENT SPECIFIES CONTAINER CONTEXT HARDWARE ACCELERATION SPECIFICATIONS FOR SINGLE TASKS WITHI…
- Which service is used by the data connection system to stream models directly into server routines?
- Which tool allows cluster admins to directly monitor real-time performance metrics of OpenShift AI routers an…
- Which kind of OpenShift deployment object handles automated horizontal scaling of model pods in a KServe serv…
Answers and explanations
Tap a question to expand the answer and the exam reasoning. Try to commit to your own pick first.
Q1. WHAT COMPONENT DEPLOYS MULTI-TENANT ROUTING ARCHITECTURES AND DECLARES INGRESS POLICIES SECURING THE OPENSHIFT AI KSERVE ROUTINES?
Answer: C. Red Hat OpenShift Serverless (Knative) and Red Hat OpenShift Service Mesh (Istio)
Knative and Istio provide the multi-tenant routing and ingress policies that secure KServe inference services. Standard OpenShift Routes or NodePorts lack the fine-grained traffic splitting needed for this architecture.
Q2. Which tool tracks a comprehensive dependency inventory to ensure pre-built Jupyter containers comply with system audits?
Answer: D. Syft or Grype container software bill of materials (SBOM) and vulnerability scanning utilities
Syft generates a Software Bill of Materials to track dependencies, while Grype scans these for vulnerabilities to ensure audit compliance. Manual commands or dashboard logs do not provide automated, comprehensive tracking required for security audits.
Q3. WHAT FIELD ASSIGNMENT SPECIFIES CONTAINER CONTEXT HARDWARE ACCELERATION SPECIFICATIONS FOR SINGLE TASKS WITHIN PIPELINE PROGRAMMATIC CODES?
Answer: C. Custom container resources request limits attributes matching nvidia.com
Hardware acceleration for GPUs is requested via custom resource limits specifying the nvidia.com domain in the container spec. Security contexts or node selectors alone do not guarantee allocation of specific hardware devices.
Q4. Which service is used by the data connection system to stream models directly into server routines?
Answer: B. Object Storage Bucket (S3/MinIO)
Object Storage Buckets like S3 are the standard for OpenShift AI data connections, allowing models to be fetched directly by serving runtimes. Git repositories are typically for notebooks or code, not large binary model storage.
Q5. Which tool allows cluster admins to directly monitor real-time performance metrics of OpenShift AI routers and pods?
Answer: C. Prometheus and Grafana
Prometheus and Grafana are the native tools for scraping and visualizing cluster performance metrics in OpenShift. The EFK logging stack only provides application logs, not time-series performance data for real-time analysis.
Q6. Which kind of OpenShift deployment object handles automated horizontal scaling of model pods in a KServe serverless system?
Answer: D. HorizontalPodAutoscaler (HPA) coupled with Knative Pod Autoscaler (KPA)
KServe on OpenShift AI relies on the Knative Pod Autoscaler, which often coordinates with the standard Horizontal Pod Autoscaler for metrics-based scaling. Standard OpenShift deployment objects are bypassed entirely by Knative Services.
Q7. What engine coordinates data science pipeline scheduling recurrences without continuous manual user web dashboard re-triggers?
Answer: B. Data Science Pipelines scheduler engine (based on Kubeflow Pipelines/Argo)
Data Science Pipelines rely on an integrated scheduler engine based on Argo and Kubeflow Pipelines to handle recurring runs. Native Kubernetes CronJobs lack the native orchestration logic required to manage complex directed acyclic graph workflows.
Q8. WHAT COMPONENT ALLOWS RECURRING SYNCHRONIZATION OF NOTEBOOKS STORED IN AN EXTERNAL REPOSITORY INTO A RUNNING DATA SCIENCE PIPELINE?
Answer: B. Elyra Git Pipeline node configuration
Elyra provides a Git pipeline node configuration that allows recurring synchronization of notebooks stored in an external repository. While Tekton executes the pipeline tasks, Elyra is the extension within JupyterLab that manages the pipeline source export.
Q9. Which sub-component handles the coordination of tasks inside an OpenShift AI Data Science Pipeline?
Answer: C. Tekton pipelines controller workflow orchestration framework systems
OpenShift AI Data Science Pipelines rely on the Tekton controller to handle task coordination and workflow orchestration. Argo CD focuses on cluster state reconciliation, not the step-by-step execution of pipeline tasks.
Q10. Which tool enables an engineer to export runtime notebook interface visualizations to reuse as pipeline assets?
Answer: D. Elyra Pipeline Visualizer workflow configurations
Elyra Pipeline Visualizer is correct because it integrates directly into Jupyter workbenches, allowing developers to convert notebook code blocks into reusable pipeline nodes. The other tools lack this native visual pipeline integration for OpenShift AI.
Q11. How do you configure cluster-wide default resources for notebook pods if a user does not specify them?
Answer: D. Editing the DataScienceCluster CRD
Editing the DataScienceCluster custom resource is correct because it contains the cluster-wide default configurations for notebook pods. The Dashboard UI lacks this specific administrator capability for setting underlying hardware defaults globally across deployments.
Q12. Which engine controls the timely dispatch and parallel processing patterns for Ray cluster compute modules in RHOAI?
Answer: C. KubeRay Operator
The KubeRay Operator manages the lifecycle and scaling of Ray clusters within OpenShift, enabling distributed compute tasks. ModelMesh is a distractor because it specifically handles standard model inference, not distributed training workloads.
Q13. Which system assures accurate metrics scraping for machine learning served routines via service monitors?
Answer: D. OpenShift User Workload Monitoring Engine
OpenShift User Workload Monitoring scrapes custom application metrics from deployed model servers using ServiceMonitors. The default platform monitoring engine specifically ignores user application workloads by design.
Q14. What logging component captures and forwards inference API transactional metadata to storage sink repositories?
Answer: A. CloudEvents Spec Logger mechanisms inside KServe architectures
KServe architecture utilizes CloudEvents specification loggers to capture and dispatch inference request metadata. While Fluentd handles generic cluster log forwarding, it is not the targeted component for payload-specific inference data.
Q15. Which repository sourcing pattern allows enforcing offline, air-gapped deployments for Red Hat OpenShift AI notebook suites?
Answer: D. Enterprise Private Mirror Registries tracking registry.redhat.io images
Private mirror registries are required for disconnected OpenShift AI deployments because they locally host the necessary container images from registry.redhat.io. Public registries like Quay or Docker Hub fail immediately in an air-gapped environment due to the lack of external network connectivity.
Q16. When configuring multi-user isolation in OpenShift AI, which network plugin feature ensures network segregation between user namespaces?
Answer: D. NetworkPolicies
NetworkPolicies ensure network segregation by isolating traffic flow between user namespaces. Components like CoreDNS handle service discovery, while Kube-Proxy manages routing rules, but neither restricts pod-to-pod communication required for strict multi-user isolation.
Q17. Which system handles incoming traffic routing to multiple separated endpoints inside an OpenShift AI Serverless environment?
Answer: A. Kourier Ingress Gateway or Istio Ingress Gateway controllers
Knative Serverless relies on the Kourier or Istio Ingress Gateway to route incoming traffic dynamically to separate endpoint revisions. Standard cluster routers like HAProxy are bypassed entirely because Knative intercepts external traffic first to manage serverless scaling.
More Red Hat Certified Specialist in OpenShift AI (EX267) drills and other practice exams are on @CertPunch. New rounds drop every few days at certpunch.com.