Tutorial 03: Azure Key Vault CSI Driver¶

This tutorial integrates the Azure Key Vault provider for Secrets Store CSI Driver so workloads can mount certificates and secrets without embedding static credentials in Kubernetes manifests.

Prerequisites¶

• Azure subscription with permission to create AKS, networking, and monitoring resources
• Azure CLI, kubectl, and a shell environment capable of exporting variables
• Existing or planned variable set for $RG, $CLUSTER_NAME, $LOCATION, and any lab-specific names
• A Log Analytics workspace resource ID stored in $WORKSPACE_ID for Container Insights validation
• Awareness that all commands use long flags only so they are easy to read and automate later

Architecture Diagram¶

flowchart TD
    POD[Workload pod] --> SA[ServiceAccount]
    SA --> FED[Federated credential]
    FED --> UAMI[User-assigned managed identity]
    UAMI --> KV[Azure Key Vault]
    KV --> CSI[Secrets Store CSI Driver]
    CSI --> POD

Step-by-Step Instructions¶

Step 1: Create Key Vault and a secret¶

az keyvault create \
    --resource-group "$RG" \
    --name "$KEYVAULT_NAME" \
    --location "$LOCATION"

az keyvault secret set \
    --vault-name "$KEYVAULT_NAME" \
    --name app-secret \
    --value "demo-value"

This step is important because it establishes the control point for create key vault and a secret. After running it, pause and verify the Azure resource state before moving on so you do not compound errors later in the lab.

Step 2: Enable the CSI driver add-on¶

az aks enable-addons \
    --resource-group "$RG" \
    --name "$CLUSTER_NAME" \
    --addons azure-keyvault-secrets-provider

This step is important because it establishes the control point for enable the csi driver add-on. After running it, pause and verify the Azure resource state before moving on so you do not compound errors later in the lab.

Step 3: Create user-assigned identity and federated credential¶

az identity create \
    --resource-group "$RG" \
    --name "$IDENTITY_NAME"

az identity federated-credential create \
    --resource-group "$RG" \
    --identity-name "$IDENTITY_NAME" \
    --name aks-csi-federation \
    --issuer "$OIDC_ISSUER" \
    --subject system:serviceaccount:workload:keyvault-reader

This step is important because it establishes the control point for create user-assigned identity and federated credential. After running it, pause and verify the Azure resource state before moving on so you do not compound errors later in the lab.

Step 4: Grant Key Vault access and apply manifests¶

az role assignment create \
    --assignee-object-id "$IDENTITY_PRINCIPAL_ID" \
    --assignee-principal-type ServicePrincipal \
    --role "Key Vault Secrets User" \
    --scope "$KEYVAULT_ID"

kubectl apply \
    --filename keyvault-serviceaccount.yaml

kubectl apply \
    --filename secretproviderclass.yaml

kubectl apply \
    --filename keyvault-pod.yaml

This step is important because it establishes the control point for grant key vault access and apply manifests. After running it, pause and verify the Azure resource state before moving on so you do not compound errors later in the lab.

Step 5: Verify mounted secret content¶

kubectl exec \
    --namespace workload \
    --stdin \
    --tty keyvault-reader \
    -- cat /mnt/secrets-store/app-secret

This step is important because it establishes the control point for verify mounted secret content. After running it, pause and verify the Azure resource state before moving on so you do not compound errors later in the lab.

Validation Steps¶

Use the following validation flow after the deployment steps complete:

• Confirm the AKS cluster and all required node pools are visible with kubectl get nodes --output wide.
• Confirm the Azure resource provisioning state is Succeeded for any new network, gateway, identity, or policy resource.
• Run at least one Container Insights query to prove telemetry is flowing before you declare the lab complete.
• Capture screenshots or exported JSON only after sanitizing identifiers such as subscription IDs or object IDs.

Example validation commands:

kubectl get pods \
    --all-namespaces \
    --output wide

az aks show \
    --resource-group "$RG" \
    --name "$CLUSTER_NAME" \
    --query "{name:name,provisioningState:provisioningState,kubernetesVersion:kubernetesVersion}" \
    --output json

az monitor log-analytics query \
    --workspace "$WORKSPACE_ID" \
    --analytics-query "KubeNodeInventory | where TimeGenerated > ago(15m) | summarize Nodes=dcount(Computer) by ClusterName" \
    --timespan "PT15M"

Cleanup Instructions¶

Delete lab resources when you are finished to avoid unnecessary spend. If the lab created shared resources that other exercises still need, remove only the lab-specific objects first.

az group delete \
    --name "$RG" \
    --yes \
    --no-wait

If you created secondary resource groups, Application Gateway, or user-assigned identities, delete those resources as part of the same cleanup workflow or document why they remain.

See Also¶

• Security
• Identity and Secrets

Sources¶

• Azure / Aks / Learn / Quick Kubernetes Deploy Cli
• Azure / Aks / Concepts Network
• Azure / Aks / Csi Secrets Store Driver
• Azure / Governance / Policy / Concepts / Policy For Kubernetes
• Azure / Azure Monitor / Containers / Container Insights Overview