Polyaxon v3 is coming →
Amazon EKS

Validate an Amazon EKS GPU cluster with Polyaxon

Amazon EKS supports several GPU operating models. EKS Auto Mode manages more of the node lifecycle and accelerator stack, while managed node groups and self-managed Karpenter leave more choices with the platform team. Polyaxon connects after that boundary is defined and keeps workload policy in queues and presets.

Outcome

An EKS capacity class validated through its actual Polyaxon queue, node image, GPU allocation mechanism, EFA path, storage, and pod identity.

Start the guideConnect a compute cluster to Polyaxon

Infrastructure responsibilities

EKS defines the cluster boundary

AWS manages the control plane. Node lifecycle, AMIs, GPU drivers, device allocation, EFA, and interruption handling depend on the selected EKS compute model.

Your platform team defines capacity

Choose instance families, zones, capacity type, node-pool policy, identity, storage, networking, and any components not managed by the selected EKS mode.

Polyaxon connects the compute cluster

Installing Polyaxon Agent (compute cluster) connects this Kubernetes cluster to the control plane. Its queues and presets route operations while Polyaxon retains run state, logs, and artifacts.

Prerequisites

  • A documented EKS compute model: Auto Mode, managed node groups, or self-managed Karpenter.
  • GPU nodes using a compatible accelerated AMI and one reviewed NVIDIA DRA driver or device-plugin path.
  • EFA-capable instances and an advertised EFA resource when the workload requires multi-node communication.
  • The cluster connected to Polyaxon by installing Polyaxon Agent (compute cluster), with a queue, storage connection, and pod identity configured for the workload namespace.

Validate the infrastructure

  1. 1

    Target the EKS capacity class

    Use platform-owned labels and a diagnostics image that matches the AMI, CUDA, libfabric, and device-allocation stack selected for this pool.

    version: 1.1kind: componentname: gpu-readinessrun:  kind: job  container:    image: nvidia/cuda:12.8.1-base-ubuntu22.04    command: ["bash", "-lc"]    args: ["nvidia-smi"]    resources:      limits:        nvidia.com/gpu: "1"

    Keep EKS placement in the selected queue or preset. This published CUDA image checks GPU visibility only; pin a tag or digest compatible with your driver. Validate EFA separately with a vetted image that includes the required libfabric and collective libraries.

  2. 2

    Submit the operation

    Run the component through the target Polyaxon project, queue, preset, and approval path.

    polyaxon run -f aws-readiness.yaml -q agent-name/queue-name
  3. 3

    Inspect the resolved EKS placement

    Confirm the operation landed on the intended capacity class and can see the GPU, EFA device when requested, storage, service DNS, and expected pod identity.

    polyaxon ops dashboard

Production checklist

Record whether AWS or your platform team owns the node lifecycle, AMI, GPU device layer, EFA integration, and interruption handling.
Treat each instance family, AMI, GPU allocation mechanism, and network topology as a separately validated capacity class.
Keep node selectors, tolerations, identity, and capacity type in platform-owned queues and presets rather than application components.
Test interruption and node rotation behavior before placing long-running training or stateful inference on reclaimable capacity.
Run NCCL tests across the intended EFA topology before accepting a distributed-training queue.

Troubleshooting

The GPU resource is missing

Confirm the node uses the intended accelerated AMI and that exactly one supported NVIDIA DRA driver or device plugin is active for that capacity model.

EFA is unavailable

Verify the instance type, attached interfaces, security group, placement, RDMA libraries, and the EFA DRA driver or device plugin used by the pool.

The connected compute cluster cannot submit workloads

Verify the Polyaxon Agent deployment, cluster registration, namespace access, pod identity, queue routing, and control-plane connectivity.

Sources

Official platform, library, model, and Polyaxon references used by this guide.

Continue

Compare AI platformsExplore GPU orchestration