NCP-GENL Production Monitoring and Reliability Practice Question
A team is deploying a large language model on NVIDIA Triton Inference Server in a production environment. They need to ensure that the model server can automatically recover from GPU failures without manual intervention. Which feature of Triton should they configure to achieve this?
⚠ Common exam trap
The trap here is assuming that Triton's internal features like instance groups or ensembles provide automatic failover, when actually orchestration-level health checks are required for recovery.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
Triton's health check endpoints and Kubernetes liveness probes
To automatically recover from GPU failures, Triton should be deployed in an orchestrated environment like Kubernetes. Triton provides health check endpoints that, when integrated with Kubernetes liveness probes, allow the orchestrator to detect an unhealthy server and restart it, possibly on a different node with a working GPU. This approach ensures minimal downtime without manual intervention.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✓
Triton's health check endpoints and Kubernetes liveness probes
Why this is correct
Triton exposes health check endpoints that Kubernetes can use with liveness probes to detect when the server is unhealthy due to GPU failure. Kubernetes can then restart the pod, potentially on a healthy GPU node, providing automatic recovery. This combination is a standard practice for achieving high availability in containerized Triton deployments.
- ✗
Instance groups with multiple GPU instances
Why it's wrong here
Instance groups allow specifying multiple instances of a model across GPUs, which can improve throughput and provide some redundancy. However, Triton does not automatically failover to a healthy instance if one GPU fails; it requires load balancing and health checks at a higher level. This configuration alone does not ensure automatic recovery.
- ✗
Model repository polling
Why it's wrong here
Model repository polling enables Triton to detect changes in the model repository and load new versions automatically. It does not address GPU failure recovery. This feature is for model updates, not for maintaining service availability during hardware faults.
- ✗
Model ensembles
Why it's wrong here
Model ensembles allow multiple models to be executed in a pipeline, but they do not provide automatic recovery from GPU failures. They are used for composing models, not for fault tolerance. This feature would not help the server recover from hardware issues without manual intervention.
About these practice questions
This NCP-GENL question is part of Courseiva's 352-question bank — original exam-style content with full explanations and wrong-answer analysis, never real exam questions or exam dumps. Learn why practice questions differ from exam dumps →
JA
Written and reviewed by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
Last reviewed September 2026 · checked against the official NVIDIA exam blueprint
This NCP-GENL practice question is part of Courseiva's free NVIDIA certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the NCP-GENL exam.