Seamlessly Deploy Cloud Native Apps - using Azure Arc and AKS

Introduction

The purpose of this document is to guide users to Deploy various Cloud Native Applications seamlessly onto an AKS cluster without explicitly containerizeing them! This also implied that now developers/archietcts coming from various domian and technology background, do not necessarily need to know te nitti-gritties of the Containers per seh for deploying their applications; but still can get the benefits of Kubernetes based of eco-sytem, its scaling, flexibility of application lifecycle management and resiliency!

We would use Azure Arc for Kubernetes as the service. This would help us to still work with our applications as-is on Cloud (App Services, Functions, Logic Apps, Message Brokers like EventGrid etc.) yet you can run it anywhere - be it on a Managed Cluster like AKS or an Un-managed cluster like CAPZ) Or K8s clusters on any other Cloud Or even K8s bare metal clusters.

This is in continuation to this [Article] where we had describe the same for deploying onto a CAPZ based Un-mnaged cluster

What the Document does

Deep-dive on Azure Arc for K8s
Create a Managed cluster - AKS
Installs all Providers and Extensions needed by Azure Arc for K8s
Creates and Deploys a Function App viz. Perform Zip built in .NetCore onto the AKS cluster
Creates and Deploys a Logic App workflow viz. NotifyZipLA onto the AKS cluster
Create an EventGrid Topic on Azure and Subscribes for any of the above services e.g. NotifyZipLA endpoint
- Push an event to EventGrid topic
- Check that the corresponding Logic App endpoint is fired!

What the Document does NOT

Deep-dive on AKS
Deep-dive on Azure Arc
Go thru the details of AKS cluster creation

Ref: Azure Arc Components

K8s Cluster can be anywhere - On-Prem, Azure or any other cloud
Azure Arc enabled K8s Service sits on Azure and can connect all these clusters as a single control plane
Azure Arc enabled agent makes outbound calls to push metadata information to Azure through Azure Arc enabled K8s Service
Azure Arc enabled agent pulls information on Agent upgrade information and Configuration from Azure through Azure Arc enabled K8s Service

Pre-requisites, Assumptions

Knowledge on Containers, Serverless Functions, Logic App - L200+
Knowledge on K8s - L200+
Knowledge on VSCode; Deploying applications through VSCode - L200+
Some knowledge on Azure tools & services viz. Azure CLI, KeyVault, VNET etc. would help

Plan

Seggregate the workload into 3 Resource Groups as per the usage
- aks-k8s-rg
  - Contains all Infrastructure components and services of the AKS cluster
    - AKS Cluster - the cluster itself
    - Virtual Network and Subnets - Hosting AKS cluster Worker Nodes
    - NSGs - for Worker Subnets
    - Key Vault - for storing any application specific Secrewts
    - Azure Container Registry (ACR) - Hosting Container images for alol applications to be deploed onto AKS cluster
- arc-k8s-workshop-rg
  - Contains Azure Arc components for K8s
    - Arc Connected Cluster for the AKS cluster
      - Establishes a connectivity to API Server running on AKS Control Plane (Master Node) and ensures Arc is managing the k8s cluster on Azure
    - CustomLocation
      - Acts as the Target location for deploying application and Data service instances on the k8s cluster.
      - Each cluster would need one or more Custom Locations
      - Each CustomLocation can host multiple app and data instances
      - To be explained later in details
    - App Service Kubernetes environment
      - Enables configuration common to apps in the custom location
      - This is required before any apps are deployed onto the cluster
- arc-services-rg
  - Contains all microservices to be deployed onto K8s Cluster
    - Zip Images
      - ZipImagesArcApp
        
        A function app to Zip images uploaded in a blob - viz. bigimageblob
        
        The zipped images are then uploaded in another blob - viz. zippedimageblob
      - NotifyZipArcLA
        
        Logic App to Read from the zippedimageblob Blob
        
        Send notification email to intended recipients with the above zipped image as an attachment
      - aks-k8s-arc-cluster-ext-eg-evg-topic
        
        Event Grid Topic which subscribes to the Logic App
        
        Send zipped image names to the NotifyZipArcLA Logic App
        
        This should be treated as a flow outside the main function flow i.e. any zippied images uploaded manually to the zippedimageblob Blob should also be sent as a notification email to the recipients
    - Validate OCR Images
      - ValidateOCRArcApp
        
        A function app to validate number plate images uploaded to a blob - viz. ocrinfoblob
        
        The images are then sent to Azure Cognitive Service for a validation of the image - using OCR API
        
        The processed response is then placed in a queue - viz. ocrinfoqueue
      - ProcessOCRImages
        
        A function app to process the OCR image response from the queue - ocrinfoqueue
        
        If the images are reasily readable by OCR API then it is moved to another Blob viz, approvedblob

Action

Working Resources

Ref: Cluster Connect

cluster-connect is a reverse-proxy agent running on the CAPZ cluster, established a session with Azure Arc Service
Any call to K8s API server would be forwarded to cluster-connect agent which passes the request to kube-aad-proxy
kube-aad-proxy reforms Azure AD authentication and if successful forwards the request to API Server for the intended cluster

Set CLI variables for easy usage and reference

tenantId="<tenant_Id>"
subscriptionId="<subscription_Id>"
arcResourceGroup="arc-k8s-workshop-rg"
aksResourceGroup="aks-k8s-rg"
arcsvcResourceGroup="arc-services-rg"
location="eastus"
clusterName="aks-k8s-arc-cluster"
version=<version>
acrName=aksk8sacr
acrId=
aksVnetName=aks-k8s-vnet
aksVnetPrefix=19.0.0.0/21
aksVnetId=
aksSubnetName=aks-k8s-subnet
aksSubnetPrefix=19.0.0.0/24
aksSubnetId=
sysNodeSize="Standard_DS3_v2"
sysNodeCount=3
maxSysPods=30
networkPlugin=azure
networkPolicy=azure
sysNodePoolName=arcsyspool
vmSetType=VirtualMachineScaleSets
addons=monitoring
aadAdminGroupID=<aadAdminGroupID>
aadTenantID==<aadTenantID?
extensionName="$clusterName-ext-appsvc"
extensionNamespace="$extensionName-ns"
customLocationName="$clusterName-custom-location"
connectedClusterName="arc-k8s-aks"

Login to Azure
```
az login --tenant $tenantId
```

Create Resource Groups as described in the Plan section

az group create -l eastus -n $aksResourceGroup
az group create -l eastus -n $arcResourceGroup
az group create -l eastus -n $arcsvcResourceGroup

Create Service Principal for CAPZ cluster

k8s-capz-sp - Name of the service principal

# Create service principal - aks-k8s-sp
az ad sp create-for-rbac --skip-assignment -n https://aks-k8s-sp

# Service Principal details
{
  "appId": "<appId>",
  "displayName": "https://aks-k8s-sp",
  "name": "<name>",
  "password": "<password>",
  "tenant": "<tenantId>"
}

Set Service Principal credentials as CLI variables
```
spAppId="<spAppId>"
spPassword="<spPassword>"
```

Create Virtual Network and corresponding Subnet hosting the AKS cluster

az network vnet create -n $aksVnetName -g $aksResourceGroup --address-prefixes $aksVnetPrefix
aksVnetId=$(az network vnet show -n $aksVnetName -g $aksResourceGroup --query="id" -o tsv)
echo $aksVnetId

az network vnet subnet create -n $aksSubnetName --vnet-name $aksVnetName -g $aksResourceGroup --address-prefixes $aksSubnetPrefix

aksSubnetId=$(az network vnet subnet show -n $aksSubnetName --vnet-name $aksVnetName -g $aksResourceGroup --query="id" -o tsv)
echo $aksSubnetId

Assign Role for Service Principal

Contributor access to the Subscription

# Create Role assignment - Contrubutor
az role assignment create --assignee $spAppId --role "Contributor" --scope "/subscriptions/$subscriptionId"

Network Contributor access to the Virtual Network viz.

az role assignment create --assignee $spAppId --role "Network Contributor" --scope $aksVnetId

Monitoring Metrics Publisher access to the Resource Group for Arc resources viz. $arcResourceGroup

arcResourceGroupId=$(az group show -n $arcResourceGroup --query="id" -o tsv)
az role assignment create --assignee $spAppId --role "Monitoring Metrics Publisher" --scope $arcResourceGroupId

Create AKS Cluster

az aks create --name $clusterName \
--resource-group $aksResourceGroup \
--kubernetes-version $version --location $location \
--vnet-subnet-id "$aksSubnetId" --enable-addons $addons \
--node-vm-size $sysNodeSize \
--node-count $sysNodeCount --max-pods $maxSysPods \
--service-principal $spAppId \
--client-secret $spPassword \
--network-plugin $networkPlugin --network-policy $networkPolicy \
--nodepool-name $sysNodePoolName --vm-set-type $vmSetType \
--generate-ssh-keys \
--enable-aad \
--aad-admin-group-object-ids $aadAdminGroupID \
--aad-tenant-id $aadTenantID \
--attach-acr $acrName

Connect to the Cluster

az aks get-credentials -g $aksResourceGroup --name $clusterName --admin --overwrite

Set Azure Arc Extension variables

connectedClusterName="arc-capz-k8s"
customLocationName="$clusterName-custom-location"
appsvcExtensionName="$clusterName-ext-appsvc"
appsvcExtensionNamespace="$clusterName-appsvc-ns"
appsvcKubeEnvironmentName="$clusterName-appsvc-kube"

Add connectedk8s extension to Azure CLI

az extension add --upgrade --yes --name connectedk8s
az extension add --upgrade --yes --name k8s-extension
az extension add --upgrade --yes --name customlocation
az extension add --upgrade --yes -n appservice-kube

Register Providers as required by Azure Arc for K8s

# Register required Providers
az provider register --namespace Microsoft.Kubernetes
az provider register --namespace Microsoft.KubernetesConfiguration
az provider register --namespace Microsoft.ExtendedLocation
az provider register --namespace Microsoft.Web

# Check Registration status of required Providers
az provider show -n Microsoft.Kubernetes -o table
az provider show -n Microsoft.KubernetesConfiguration -o table
az provider show -n Microsoft.ExtendedLocation -o table
az provider show -n Microsoft.Web -o table

Deploy App Service Extension on the AKS cluster

az k8s-extension create \
    --resource-group $arcResourceGroup \
    --name $extensionName \
    --cluster-type connectedClusters \
    --cluster-name $connectedClusterName \
    --extension-type 'Microsoft.Web.Appservice' \
    --release-train stable \
    --auto-upgrade-minor-version true \
    --scope cluster \
    --release-namespace $extensionNamespace \
    --configuration-settings "Microsoft.CustomLocation.ServiceAccount=default" \
    --configuration-settings "appsNamespace=${extensionNamespace}" \
    --configuration-settings "clusterName=${extensionName}" \
    --configuration-settings "keda.enabled=true" \
    --configuration-settings "buildService.storageClassName=default" \
    --configuration-settings "buildService.storageAccessMode=ReadWriteOnce" \
    --configuration-settings "customConfigMap=${extensionNamespace}/kube-environment-config" \
    --configuration-settings "envoy.annotations.service.beta.kubernetes.io/azure-load-balancer-resource-group=${aksResourceGroup}"

Check the status of the App Service Extension creation
```
az k8s-extension show -c $connectedClusterName --cluster-type connectedClusters   -n $appsvcExtensionName -g $arcResourceGroupName
```
Ref: K8s-Extensions
- Different extensions for different Service types to be deployed on the K8s Cluster
  - App Service Extension - Includes Web Apps/API Apps, Logic Apps, Function Apps
  - EventGrid Extension - Supports EventGrid Topic deployment
  - Data Service Extension - Supports deployment of SQL MI
- config-agent - Keeps track fo new or updated extension resources. Multiple extensions can be added on the same cluster with different versions
- extension-manager - Pulls Helm chart from Container Registry and deploys onto the K8s cluster

Retrieve the ExtensionId to be used in subsequent steps

extensionId=$(az k8s-extension show \
  --cluster-type connectedClusters \
  --cluster-name $connectedClusterName \
  --resource-group $arcResourceGroupName \
  --name $appsvcExtensionName \
  --query id \
  --output tsv)
echo $extensionId

# Check Pods created in teh extensionNamespace
k get po -n $appsvcExtensionNamespace

Create CustomLocation with Azure Arc Connected Cluster
- Every Application Services or Data Services would be deployed in the CustomLocation rather than an Azure Region/Location
- This would ensure various types of application and data services can run together in same Arc Enabled Cluster
```
connectedClusterId=$(az connectedk8s show --resource-group $arcResourceGroupName --name $connectedClusterName --query id --output tsv)
echo $connectedClusterId

az customlocation create \
--resource-group $arcResourceGroupName \
--name $customLocationName \
--host-resource-id $connectedClusterId \
--namespace $appsvcExtensionNamespace \
--cluster-extension-ids $extensionId
```
  Ref: Custom Location
  - Target Location for all services
  - As evident from the diagram - it is a layer around Azure Arc enabled K8s
  - Ideally one custom location per k8s cluster
  - Every allowed services on Azure Arc would be deployed to this Custom Location; instead of an Azure region
  - Creates K8s ClusterRoleBindings and RoleBindings
  - This will ensure that Azure Arc enabled K8s can deploy necessary resources viz. pods, services, replicates, statefulsets etc.) for the deployed application services i.e. API Apps, FunctionApps, Logic Apps etc.

Check the status of CustomLocation creation process

az customlocation show --resource-group $arcResourceGroupName --name $customLocationName

Retrieve the CustomLocationId to be used in subsequent steps

customLocationId=$(az customlocation show \
--resource-group $arcResourceGroupName \
--name $customLocationName \
--query id \
--output tsv)
echo $customLocationId

Create App Service Kube Environment for the above CustomLocation
- This is a collection of all App Service Plans and App Services
- Please note that this is only needed for Application Services; for Data Services thsi would be performed by Data Controllers for Arc
```
az appservice kube create \
--resource-group $arcResourceGroupName \
--name $appsvcKubeEnvironmentName \
--custom-location $customLocationId
```

Check Kube Environment creation process

az appservice kube show \
--resource-group $arcResourceGroupName \
--name $appsvcKubeEnvironmentName

Use Case Flow

Zip Images

This is what we would like to achieve in this exercise on the Custom Location created above:

Deploy a .NetCore Http triggerred Function App - ZipImagesArcApp
Deploy a Locally developed Blob triggered Logic App - NotifyZipArcLA Logic App
Deploy an EventGrid topic aks-k8s-arc-cluster-ext-eg-evg-topic
Test
- Upload Images to the Blob -viz. bigimgeblob
- Call Http triggererd function Url of ZipImagesArcApp passing the arrya iof image names in the
- Push an event to EventGrid topic from PostMan or any other Rest client
- Check that the corresponding Subscription Endpoint (e.g. NotifyZipArcLA) is fired!

Function App

This execise uses a simple Http Triggerred Azure Function in .NetCore - ZipImagesArcApp for this purpose. One can use any type of Azure Function of their choice
Visual Studio Code or Visual Studio both have easy integration with Azure Resource management. Any other IDE with appropriate plugins can be used as well. This exercise would use VSCode as an option
Open Function App root folder in Visual Studio Code
Right Click and Deploy to Function App. Please note one can create the Function App in the portal and then manage deployment from VSCode
VSCode would ask for a new App to be Created Or Deploy on an existing one
The Target Location step is extremely important - should be the CustomLocation created in earlier steps
Once the steps are completed, comeback to Azure CLI
Check Deployments and/or Pods of the App Service Namespace in the K8s cluster. All Pods should be in the running state
Go to Azure Portal and Check the Function App resource; in the Overview blade it will show up the Web API access URL
Check the URL in the browser; use Postman or any REST client to call to test different paths of the Function App

Logic App

This execise uses a simple Http Triggerred Logic App Created Locally - NotifyZipArcLA for this purpose
Few points to note here on the choice of Creation path to Azure and subsequent Deployment onto K8s cluster
- This Logic App type would be Standard and Stateful which is actually a Single Tenant Logic App; rather than the Consumption type Logic App which is Multi-Tenant Logic App
- Currently the best way to achieve a seamless experiene end-to-end is to Create and Deploy Logic App Standard, Stateful type from Visual Studio Code itself
- Not all triggers are available for Standard mode as of now
- This exercise uses a simple Blob trigger for demonstration
Visual Studio Code or Visual Studio both have easy integration with Azure Resource management. Any other IDE with appropriate plugins can be used as well. This exercise would use VSCode as an option
Open Logic App root folder in Visual Studio Code
Right Click and Deploy to Logic App
VSCode would ask for a new App to be Created Or Deploy on an existing one
The Target Location step is extremely important - should be the CustomLocation created in earlier steps
Once the steps are completed, comeback to Azure CLI
Check Deployments and/or Pods of the App Service Namespace in the K8s cluster. All Pods should be in the running state
Uplaod some blob images and check that the Logic App worlflow gets triggered

Event Grid

This exercise uses a simple Event Grid Topic - aks-k8s-arc-cluster-ext-eg-evg-topic for this purpose

Set local varibales in Azure CLI

evgExtensionName="$clusterName-ext-eg"
evgExtensionNamespace="$clusterName-evg-ns"
evgTopicName="$clusterName-egt"
evgSubscriptionName="$clusterName-evg-sub"

Create Event Grid extension for Azure Arc on K8s

This step is easy to be done through portal as there are many config options - https://docs.microsoft.com/en-us/azure/event-grid/kubernetes/install-k8s-extension

Check the event Grid extension creation process

az k8s-extension show -c $connectedClusterName --cluster-type connectedClusters \
-n $evgExtensionName -g $arcK8sResourceGroup

Arc Enabled Services

Make sure to Select the EventGrid service from the dropdown before creating the EventGrid Topic - otherwise Topic creation would fail with Forbidden error; as it would not recognize the target deployment Location - which should be the Custom Location in the case
Visual Studio Code or Visual Studio both DONOT have integration with Azure Arc flavour for EventGrid as of now. So, creating the Topic in portal is the only option as of now
The Target Location step is extremely important - should be the CustomLocation created in earlier steps

Check Topic details

topicId=$(az eventgrid topic show --name $evgTopicName --resource-group $arcSvcResourceGroup --query id -o tsv)
echo $topicId

Create Event Subscription for the above topic

az eventgrid event-subscription create --name $evgSubscriptionName --source-resource-id $topicId \
--endpoint <event_subscription_endpoint>

 # e.g. function app endpoint that we had created earlier

Check Deployments and/or Pods of the EventGrid Namespace in the K8s cluster. All Pods should be in the running state
```
k-capz get po -n $eventGridnamespace
```

Get EventGrid Endpoint details

eventGridEndpoint=$(az eventgrid topic show --name $evgTopicName -g $arcSvcResourceGroup --query "endpoint" --output tsv)

Get EventGrid Key details

eventGridKey=$(az eventgrid topic key list --name $evgTopicName -g $arcSvcResourceGroup --query "key1" --output tsv)

Following piece of code would be executed when a POST call is made for the /upload endpoint

Send the EventGrid Endpoint in Http request header as key - ep; e.g. ep = ""
Send the EventGrid Key in Http request header as key - key; e.g. key = ""
Only supported Event Schema as of now - CloudEvent Schema; EventGrid Schema will not work currently!

express.post("/upload", async (request, response) =>
{

    try
    {

        const requestHeaders = request.headers;
        const evgHeaders =
        {

            "Content-Type": "application/cloudevents-batch+json",
            "aeg-sas-key": requestHeaders.key

        };

        const config =
        {

            headers: evgHeaders

        };

        const data =
        { 
            "specversion": "1.0",
            "type" : "orderCreated",
            "source": "myCompanyName/us/webCommerceChannel/myOnlineCommerceSiteBrandName",
            "id" : "eventId-n",
            "time" : "2020-12-25T20:54:07+00:00",
            "subject" : "account/acct-123224/order/o-123456",
            "dataSchema" : "1.0",
            "data" : {
               "orderId" : "123892",
               "orderType" : "PO",
               "reference" : "https://www.myCompanyName.com/orders/123"
              }
        };

        const result = await Axios.post(requestHeaders.ep, [data], config);
        console.log(result);
        response.send("PostAPIApp - " + result.data);
        
    }
    catch(exception)
    {

        response.send("Error - " + exception.message);

    }

This in-turn calls the subscription endpoint; check if the PostMessageApp Function being called

Additional Points

The same exercise would work for AKS cluster also almost as-is
PersistentVolume Step through Helm chart would NOT be needed. AKS being a mangled Cluster - it would come-up with a set of Storage Classes
Repository for this Source Code

References

Azure Arc Enabled K8s:

AppServices on K8s

EventGrid on K8s

DataServices on K8s

Name		Name	Last commit message	Last commit date
Latest commit History 85 Commits
Assets		Assets
Deployments		Deployments
Functions		Functions
LogicApps		LogicApps
Misc		Misc
.gitignore		.gitignore
README-DataController.md		README-DataController.md
README.md		README.md

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

Seamlessly Deploy Cloud Native Apps - using Azure Arc and AKS

Introduction

What the Document does

What the Document does NOT

Pre-requisites, Assumptions

Plan

Seggregate the workload into 3 Resource Groups as per the usage

Arc Connected Cluster for the AKS cluster

CustomLocation

App Service Kubernetes environment

Action

Working Resources

Use Case Flow

Zip Images

Function App

Logic App

Event Grid

Additional Points

References

About

Releases

Packages

Contributors 2

Languages

monojit18/AKS-Arc

Folders and files

Latest commit

History

Repository files navigation

Seamlessly Deploy Cloud Native Apps - using Azure Arc and AKS

Introduction

What the Document does

What the Document does NOT

Pre-requisites, Assumptions

Plan

Seggregate the workload into 3 Resource Groups as per the usage

Arc Connected Cluster for the AKS cluster

CustomLocation

App Service Kubernetes environment

Action

Working Resources

Use Case Flow

Zip Images

Function App

Logic App

Event Grid

Additional Points

References

About

Resources

Stars

Watchers

Forks

Releases

Packages 0

Contributors 2

Languages

Packages