Deploy kagent

Running Agentic / Platform Engineering Assistant / lab 2 of 7View on GitHub

Installation

Below you will find the installation of kagent for Llama as your first Model used (you can add more after installation)

Helm

CRDs

helm install kagent-crds oci://ghcr.io/kagent-dev/kagent/helm/kagent-crds \
    --namespace kagent \
    --create-namespace

With Llama

The below contains the flag to give the kagent UI a public IP so you can reach it that way instead of doing a port-forward. However, if you’re running kagent locally or don’t want to create a load balancer, you can just remove the --set ui.service.type=LoadBalancer part of the installation below.

helm upgrade --install kagent oci://ghcr.io/kagent-dev/kagent/helm/kagent \
    --namespace kagent \
    --set providers.default=ollama \
    --set ui.service.type=LoadBalancer

Dashboard Setup

  1. With an LB
kubectl get svc -n kagent
  1. Retrieve the public IP of the kagent-ui LB

Without an LB

kubectl port-forward svc/kagent-ui -n kagent 8080:8080
  1. Access the kagent setup and create your own agent via the UI