Skip to main content

How do I connect Harness to my environment?

Harness performs deployments using a service named the Harness Delegate. You install the Delegate in your environment and it makes a secure, outbound connection to Harness.

When you run a Harness pipeline, the Delegate pulls in the artifacts and manifests/specs needed for deployment, and then runs deployment steps in your environment.

To install a Delegate, you copy the Helm commands provided by Harness and run them in your target environment. Next, the Delegate will register with Harness. That's it.

Watch this video to see the steps to install your delegate

If you don't have Helm installed, see these Helm install instructions.

What size cluster do I need?​

The Delegate only requires 1 replica with 2GB MEM and 0.5 CPU. But since you are also installing your service or the Harness sample service, use a cluster with at least 2 replicas with 4GB MEM and 1 CPU. If your service is large, use a larger cluster.

Cluster provisioning​

If you don't have a cluster already, connect to your platform and use the following commands to set one up before installing the Delegate.

Minikube (local option)​

If you want to use Minikube, use Minikube minimum version v1.22.0 or later installed locally.

minikube start --memory=4g --cpus=1 --nodes=2

After running this command, you can verify that the cluster is running by running the following command:

kubectl get nodes

Docker Desktop (local option)​

To set up a Kubernetes cluster with 2 replicas, 4GB of memory, and 1 CPU in Docker Desktop, you can follow these steps:

  1. Open Docker Desktop and go to the Settings menu.
  2. Select the Kubernetes tab.
  3. Enable Kubernetes by checking the Enable Kubernetes checkbox.
  4. Set the number of replicas to 2 by adjusting the Replicas slider.
  5. Set the amount of memory to 4GB and the number of CPUs to 1 by adjusting the Memory and CPU sliders.
  6. Select the Apply & Restart button to apply the changes and restart Docker Desktop.

After Docker Desktop restarts, you can verify that the cluster is running by running the following command:

kubectl get nodes

Google GKE provisioning​

Replace ZONE with your GCP region, for example us-central1-c:

gcloud container clusters create [CLUSTER-NAME] --num-nodes=2 --machine-type=[MACHINE-TYPE] --disk-size=10GB --zone=[ZONE]

For example:

gcloud container clusters create my-cluster --num-nodes=2 --machine-type=n1-standard-1 --disk-size=10GB --zone=us-central1-a

After running this command, you can verify that the cluster is running by running the following command:

kubectl get nodes

Azure AKS provisioning​

Replace myResourceGroup with your AKS resource group:

az aks create -g myResourceGroup -n myAKSCluster --enable-managed-identity --node-count 2 --enable-addons monitoring --enable-msi-auth-for-monitoring --generate-ssh-keys

After running this command, you can verify that the cluster is running by running the following command:

kubectl get nodes

AWS EKS provisioning​

eksctl create cluster --name=my-cluster --version=1.21 --nodegroup-name=my-nodegroup --node-type=t3.small --nodes-min=2 --nodes-max=2 --node-volume-size=20 --region=us-west-2

After running this command, you can verify that the cluster is running by running the following command:

kubectl get nodes