Skip to content

Latest commit

 

History

History
196 lines (137 loc) · 7.62 KB

Readme.md

File metadata and controls

196 lines (137 loc) · 7.62 KB
Raw

Application Insights for Kubernetes on WebAPI

This page is a walk-through to enable Application Insights for Kubernetes in WebAPI projects, including traditional WebAPI (controller-based) and minimal WebAPIs. The example code is separated in WebAPI and MinimalAPI respectively.

To learn more about minimal API, please refer to the Minimal APIs overview.

Prerequisite

  • .NET 6 SDK - https://dot.net

    • Verify dotnet version. For example: 
      dotnet --version
6.0.302

    You might have a slightly different patch version and that should be fine.

  • A Kubernetes Cluster that you can manage with kubectl.

    • If you don't have any, there are several options:

    • Verify that the credential is properly set up for kubectl, for example:

      kubectl get nodes
NAME                        STATUS   ROLES           AGE    VERSION
aks-nodepool1-10984277-0    Ready    agent           5d8h   v1.24.1

  • A container image registry

    • The image built will be pushed into an image registry. Dockerhub is used in this example. It should work with any image registry.

Prepare the project

  1. Create the project from templates

    • For Control-Based WebAPI:

      dotnet new webapi           # Create a control-based webapi from the template

    • For Minimal WebAPI

      dotnet new web             # Create a minimal webapi from the template

  2. Add NuGet packages

    dotnet add package Microsoft.ApplicationInsights.AspNetCore
dotnet add package Microsoft.ApplicationInsights.Kubernetes --prerelease

Enable Application Insights for Kubernetes

To enable Application Insights for Kubernetes, the services need to be registered:

...
// Enable application insights
builder.Services.AddApplicationInsightsTelemetry();

// Enable application insights for Kubernetes
builder.Services.AddApplicationInsightsKubernetesEnricher(diagnosticLogLevel: LogLevel.Information);
...

See a full example for WebAPI or Minimal WebAPI.

Prepare the image

  1. Get the dockerfile ready. For example, refer to this for WebAPI or this for Minimal Web API, remember to update the ENTRYPOINT to align with your assembly name. For more about containerizing an ASP.NET application, refer to Dockerize an ASP.NET Core application.

  2. Setup some variables for the convenience

    $imageName='ai-k8s-web-api'                     # Choose your own tag name
$imageVersion='1.0.0'                           # Update the version of the container accordingly
$testAppName='test-ai-k8s-web-api'              # Used as the container instance name for local testing
$imageRegistryAccount='your-registry-account'   # Used to push the image

    Note: In this example, we use PowerShell. Tweak it accordingly in other shells. Also, you don't have to have those variables if you prefer to use the value directly below.

  3. Build the image

    docker build -t "$imageName`:$imageVersion" .   # For PowerShell, you need to escape the colon(:) with (`).
docker container rm $testAppName -f             # Making sure any existing container with the same name is deleted
docker run -d -p 8080:80 --name $testAppName "$imageName`:$imageVersion"

  4. Check the logs

    docker logs $testAppName

    When you see this error at the beginning of the logs, it means the container is ready to be put into a Kubernetes cluster:

    [Error] [2022-08-12T23:56:08.9973664Z] Application is not running inside a Kubernetes cluster.

  5. Remove the running container

    docker container rm $testAppName -f

  6. Tag and push the image

    docker tag "$imageName`:$imageVersion" "$imageRegistryAccount/$imageName`:$imageVersion"
docker push "$imageRegistryAccount/$imageName`:$imageVersion"

Deploy to Kubernetes Cluster

Now that the image is in the container registry, it is time to deploy it to Kubernetes.

  1. Deploy to a dedicated namespace

    You could use the default namespace, but it is recommended to put the test application in a dedicated one, for example, 'ai-k8s-demo'. To deploy a namespace, use content in k8s-namespace.yaml:

    kubectl create -f ..\k8s-namespace.yaml # tweak the path accordingly

  2. Setup proper role binding for RBAC-enabled clusters

    If you have RBAC enabled for your cluster, permissions need to be granted to the service account to access the cluster info for telemetries. Refer to Configure RBAC permissions for details.

    For example, deploy a role assignment in the namespace of ai-k8s-demo by using sa-role.yaml:

    kubectl create -f ..\..\docs\sa-role.yaml # tweak the path accordingly

    Notes: Check the namespace for the service account in the yaml file.

  3. Deploy the application

    Create a Kubernetes deployment file. Reference k8s.yaml as an example, pay attention to namespace, image, and environment variables, making sure they are properly set up.

    Then run the following command to deploy the app:

    kubectl create -f k8s.yaml

    To learn more about Kubernetes deployment, please refer to Deployments.

  4. Check out the pod deployed:

    kubectl get pods --namespace ai-k8s-demo    # Or whatever your namespace of choice.

    And you will see something like this:

    PS > kubectl get pods --namespace ai-k8s-demo
NAME                                             READY   STATUS    RESTARTS   AGE
some-other-pod-that-has-run-for-10-days-pmfdc    1/1     Running   0          10d
ai-k8s-web-api-deployment-97d688b46-7cxs5        1/1     Running   0          2m58s           <==== created ~3 minutes ago

    Check out the logs, for example:

    PS > kubectl logs ai-k8s-web-api-deployment-97d688b46-7cxs5 --namespace ai-k8s-demo
[Information] [2022-08-23T21:32:01.0135029Z] [CGroupContainerIdProvider] Got container id: 41a149977b4ea424947d54c59e7f63d32664e19503adda762004250f72db7687
[Information] [2022-08-23T21:32:01.2737318Z] Found pod by name providers: ai-k8s-web-api-deployment-97d688b46-7cxs5
info: Microsoft.Hosting.Lifetime[14]
    Now listening on: http://[::]:80
info: Microsoft.Hosting.Lifetime[0]
    Application started. Press Ctrl+C to shut down.
info: Microsoft.Hosting.Lifetime[0]
    Hosting environment: Production
info: Microsoft.Hosting.Lifetime[0]
    Content root path: /app/

  5. Test the endpoint

    One way to hit the endpoint is by port forwarding. Check out the example in Deploy the application in Kubernetes, looking for "Port forward for testing" section specifically.

  6. Delete the cluster after the test

    PS > kubectl delete -f k8s.yaml
deployment.apps "ai-k8s-web-api-deployment" deleted