This example walks through the steps to deploy an ASP.NET Core 2.1 MVC application to a Kubernetes cluster with Application Insights for Kubernetes on. The steps also applies for ASP.NET Core 2.2 applications.

A simple cluster role sample yaml is also included to describe how to make it work in a Role-based access control(RBAC) enabled clusters.

Tip: Read this for more information about RBAC.

Note: This is an example that does NOT follow all best practices, including security-related best practices. E.g. Application Insights instrumentation key is not adequately protected (it should be deployed as a secret).

• .NETCore SDK 2.1.300 or above

  • .NET Core SDK is required in this example. Go to https://dot.net to download the latest SDK. Make sure you have 2.1.300 or above:

    dotnet --version
    2.1.301

• A Kubernetes Cluster that you can manage with kubectl.

  • If you don't have any, an easy way is to go to Azure AKS to get a managed cluster. Verify that the credential is properly set for kubectl to work:

    user@user-pc:~$ kubectl get nodes
    NAME                       STATUS    ROLES     AGE       VERSION
    aks-nodepool1-10984277-0   Ready     agent     17d       v1.9.9
    aks-nodepool1-10984277-1   Ready     agent     17d       v1.9.9
    aks-nodepool1-10984277-2   Ready     agent     17d       v1.9.9
    user@user-pc:~$

• A container image repository

  • The image built will be pushed into an image repository. Dockerhub is used in this example.

• Let's start by creating an ASP.NET Core MVC applicaiton:

dotnet new mvc

• Add the NuGet Packages:

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

• Enable Application Insights and Application Insights for Kubernetes in Startup.cs:

public void ConfigureServices(IServiceCollection services)
{
    ...
    services.AddApplicationInsightsTelemetry();             // Enables Application Insights
    services.AddApplicationInsightsKubernetesEnricher();    // Enables Application Insights for Kubernetes.
    ...
}

• It is optional but recommended to update the base images:

docker pull mcr.microsoft.com/dotnet/core/sdk:2.1
docker pull mcr.microsoft.com/dotnet/core/aspnet:2.1

• Add Dockerfile to the project folder. Build the docker container (dockeraccount/aik8sbasic_rbac, for example) using Dockerfile and upload it to an image registry.

docker build . -t dockeraccount/aik8sbasic_rbac:latest
docker push dockeraccount/aik8sbasic_rbac:latest

• If the cluster is RBAC enabled, the service account used will need to bind to proper cluster role so that the application can fetch Kubernetes related properties.

In saRole.yaml, a cluster role named appinsights-k8s-property-reader is created and then bind to the default service account. Permissions needed are listed in the resources property. To deploy it, update the value for the namespace and then:

kubectl create -f k8s/saRole.yaml

• Create the Kubernetes spec for the deployment and the service. Referencing k8s.yaml. Please update the variable of APPINSIGHTS_INSTRUMENTATIONKEY to your own application insights instrumentation key.

Deploy it:

kubectl create -f k8s/k8s.yaml

Once properly set up, your telemetry data will all be decorated with Kubernetes properties on it:

• Troubleshoot Application Insights for Kubernetes

• Enable Application Insights Profiler to optimize the performance for your application.