Sourcegraph with Kubernetes on Amazon EKS

Amazon EKS is Amazon's managed Kubernetes offering, similar to how Google Cloud offers managed Kubernetes clusters (GKE).

If your preferred cloud provider is Amazon, we strongly recommend using EKS instead of plain EC2. By using EKS, you will not need to manage your own Kubernetes control plane (complex). Instead, Amazon will provide it for you and you will only be responsible for managing the NodeGroups and the Sourcegraph deployment running on the Kubernetes cluster.

This guide will help you create a simple Kuberentes cluster using EKS, for more information or other advanced use-cases, please check Amazon EKS documentation.

Requirements

The easiest way to get started with Amazon EKS is using the eksctl. This tool will create all the necesary resources to bootstrap a simple Kuberentes cluster in AWS. You can find an extensive guide on using eksctl with Amazon EKS in the EKS Getting Starrted guide.

Before moving forward, you will need:

  • awscli: The command line tool provided by Amazon to work with AWS resources. Installation guide and Configuration guide.
  • kubectl: The command line tool used for working with Kuberentes clusters and resources. Installation guide.
  • eksctl: A tool provided by Waveworks for configuring and creating EKS clusters in Amazon. Installation guide.
  • IAM permissions: The user that will be performing the installtion must have permissions to work with Amazon EKS IAM roles and service linked roles, AWS CloudFormation, and a VPC and related resources.

Through this guide, we will be using a number of parameters starting with $ that you will need to replace with your desired values:

  • $REGION: The AWS region to use for all resources in this tutorial.
  • $KEY_NAME: The name of the SSH key-pair in AWS used to access instances.
  • $CLUSTER_NAME: A name to be given to the Amazon EKS Cluster.
  • $NODE_TYPE: The instance type that will be used by the cluster NodeGroup.
  • $NODE_MAX: The maximum number of nodes in your cluster.
  • $NODE_MIN: The minimum number of nodes in your cluster.

Sizing the cluster

Before getting started, we need to identify the size for our initial cluster. You can use the following chart as reference:

Users $NODE_TYPE $NODE_MIN $NODE_MAX Cost est.
10-500 m5.4xlarge 3 6 $59-118/day
500-2000 m5.4xlarge 6 10 $118-195/day

Note: You can modify these values later on to scale up/down the number of worker nodes using the eksctl command line. For more information please the eksctl documentation

Create an EC2 key-pair

Note: If you already have an existing keypair you can skip this step and reuse your existing key

Creating an EC2 keypair will allow you to access the Kuberentes nodes after the setup is complete, which might be required in the future to perform troublshooting or administration tasks.

To create a key-pair in use the following command:

aws ec2 create-key-pair --region $REGION --key-name $KEY_NAME

Create an Amazon EKS Cluster

We will leverage eksctl to create our cluster, as it automates all the steps necesary to get our EKS Cluster and its NodeGroup working.

Note: This command might take 10-30 minutes to finish.

eksctl create cluster \
  --name $CLUSTER_NAME \
  --region $REGION \
  --with-oidc \
  --ssh-access \
  --ssh-public-key $KEY_NAME \
  --managed \
  --node-type $NODE_TYPE --nodes-min $NODE_MIN --nodes $NODE_MIN --nodes-max $NODE_MAX

Access to the cluster

Before accessing the cluster, we need to configure kubectl to connect to the correct cluster, we can do this with eksctl:

eksctl --region $REGION utils write-kubeconfig --cluster $CLUSTER_NAME

At this point, running kubectl get svc should show something like:

$ kubectl get svc
NAME         TYPE        CLUSTER-IP   EXTERNAL-IP   PORT(S)   AGE
kubernetes   ClusterIP   172.20.0.1   <none>        443/TCP   4m

Important: If kubectl commands prompt you for username/password, make sure that kubectl version reports a client version of v1.10+. Older versions of kubectl do not work with the authentication configuration provided by Amazon EKS.

Deploy the Kubernetes Web UI Dashboard (optional)

See Tutorial: Deploy the Kubernetes Dashboard.

Deploy Sourcegraph! 🎉

Your Kubernetes cluster is now all set up and running!

Luckily, deploying Sourcegraph on your cluster is much easier and quicker than the above steps. :)

Follow our installation documentation to continue.