Sourcegraph with Kubernetes on Azure

WARNING: This guide applies exclusively to a Kubernetes deployment without Helm. If you have not deployed Sourcegraph yet, it is higly recommended to use Helm as it simplifies the configuration and greatly simplifies the later upgrade process. See our guidance on using Helm to deploy to Azure AKS.

Install the Azure CLI tool and log in:

SHELL
az login

Sourcegraph on Kubernetes requires at least 16 cores in the DSv3 family in the Azure location of your choice (e.g. eastus), so make sure you have enough available (if not, request a quota increase):

SHELL
$ az vm list-usage -l eastus -o table
Name                                CurrentValue    Limit
--------------------------------  --------------  -------
...
Standard DSv3 Family vCPUs                     0       32
...

Ensure that these Azure service providers are enabled:

SHELL
az provider register -n Microsoft.Network
az provider register -n Microsoft.Storage
az provider register -n Microsoft.Compute
az provider register -n Microsoft.ContainerService

Create a resource group:

SHELL
az group create --name sourcegraphResourceGroup --location eastus

Create a cluster:

SHELL
az aks create --resource-group sourcegraphResourceGroup --name sourcegraphCluster --node-count 1 --generate-ssh-keys --node-vm-size Standard_D16s_v3

Connect to the cluster for future kubectl commands:

SHELL
az aks get-credentials --resource-group sourcegraphResourceGroup --name sourcegraphCluster

Follow the Sourcegraph cluster installation instructions with storageClass set to managed-premium in config.json:

DIFF
-    "storageClass": "default"
+    "storageClass": "managed-premium"

You can see if the pods are ready and check for installation problems through the Kubernetes dashboard:

SHELL
az aks browse --resource-group sourcegraphResourceGroup --name sourcegraphCluster

Set up a load balancer to make the main web server accessible over the network to external users:

SHELL
kubectl expose deployment sourcegraph-frontend --type=LoadBalancer --name=sourcegraphloadbalancer --port=80 --target-port=3080