EVA Vision Installation

This guide explains how to install EVA Vision, the module responsible for real-time vision analysis of video streams, in a Kubernetes cluster.

🛠️ Prerequisites

Kubernetes cluster
kubectl
Helm
AWS CLI and ECR permissions when pulling images from AWS ECR

aws configure
# Enter your AWS Access Key ID and Secret Access Key.

🚀 Install EVA Vision

Step 1: Register the Helm Repository

helm repo add eva-vision <https://raw.githubusercontent.com/mellerikat/eva-vision/chartmuseum/>
helm repo update

Step 2: Prepare the Default Values File

helm show values eva-vision/eva-vision > values.yaml

Step 3: Update Environment-Specific Settings

Modify values.yaml or create a separate override for cloud/AWS or on-premises/k3s.

Category	Parameter	Description	On-premises	Cloud/AWS
Image	`image.tag`	Docker image tag	`v2`	`v2`
Resources	`resources.limits`	GPUs allocated per pod	`nvidia.com/gpu: 1`	`nvidia.com/gpu: 1`
Storage	`persistence.storageClass`	Storage class for logs	`local-path`	`eva-agent-sc-bs`
Node placement	`affinity`	GPU-node placement	Empty	Specify `g6.2xlarge`, etc.
Model management	`marName`	AI model `.mar` files	`Owl-v2, ig, OmDet...`	`Owl-v2, ig, OmDet...`
Workers	`maxWorkers`	Maximum workers	`1`	`1`
Inference	`maxBatchDelay`	Maximum batching delay	`150ms`	`150ms`

AWS EKS Override Example (`eva-vision-aws.yaml`)

persistence:
  storageClass: "eva-agent-sc-bs"
  size: 30Gi

affinity:
  nodeAffinity:
    requiredDuringSchedulingIgnoredDuringExecution:
      nodeSelectorTerms:
      - matchExpressions:
        - key: node.kubernetes.io/instance-type
          operator: In
          values:
          - g6.2xlarge

Local k3s Override Example (`eva-vision-k3s.yaml`)

persistence:
  storageClass: "local-path"
  size: 20Gi

affinity: {}