How-To: Patch Kubernetes resources using PodSpec

This how-to guide will provide an overview of how to:

• Patch Radius-created Kubernetes resources using PodSpec definitions

Prerequisites

• rad CLI
• Radius initialized with rad init
• kubectl

Step 1: Define a container

Begin by creating a file named app.bicep with a Radius container:

extension radius

@description('Specifies the environment for resources.')
param environment string

resource app 'Applications.Core/applications@2023-10-01-preview' = {
  name: 'demo'
  properties: {
    environment: environment
  }
}

resource demo 'Applications.Core/containers@2023-10-01-preview' = {
  name: 'demo'
  properties: {
    application: app.id
    container: {
      image: 'ghcr.io/radius-project/samples/demo:latest'
      ports: {
        web: {
          containerPort: 3000
        }
      }
    }
  }
}

Step 2: Deploy the app and container

1. Deploy and run your app:

   rad run ./app.bicep -a demo
   

   Once the deployment completes successfully, you should see the following confirmation message along with some system logs:

   Deployment Complete
   
   Resources:
      demo            Applications.Core/applications
      demo            Applications.Core/containers
   
   Starting log stream...
   
   + demo-df76d886c-sngm8 › demo
   demo-df76d886c-sngm8 demo Using in-memory store: no connection string found
   demo-df76d886c-sngm8 demo Server is running at http://localhost:3000
   demo-df76d886c-sngm8 demo [port-forward] connected from localhost:3000 -> ::3000
   

   Access the application by opening http://localhost:3000 in a browser, where you should see the demo app:

   

   
   When you’re ready to move on to the next step, use CTRL + C to exit the command.

2. Run the command below, which will list the pods in your Kubernetes cluster, using the -o flag to specify the relevant information to output:

   kubectl get pods -A -l app.kubernetes.io/name=demo -o custom-columns=POD:.metadata.name,STATUS:.status.phase,CONTAINER_NAMES:spec.containers[:].name,CONTAINER_IMAGES:spec.containers[:].image
   

   You should see output confirming that a single container named demo was deployed and is running in your pod, similar to the following:

   POD                    STATUS    CONTAINER_NAMES   CONTAINER_IMAGES
   demo-df76d886c-9p4gv   Running   demo              ghcr.io/radius-project/samples/demo:latest
   

Step 3: Add a PodSpec to the container definition

Add the following runtimes configuration to the container definition in your app.bicep file. This allows you to punch through the Radius abstraction and directly apply any part of the Kubernetes PodSpec. In this example you’re adding an additional sidecar container:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39

extension radius

@description('Specifies the environment for resources.')
param environment string

resource app 'Applications.Core/applications@2023-10-01-preview' = {
  name: 'demo'
  properties: {
    environment: environment
  }
}

resource demo 'Applications.Core/containers@2023-10-01-preview' = {
  name: 'demo'
  properties: {
    application: app.id
    container: {
      image: 'ghcr.io/radius-project/samples/demo:latest'
      ports: {
        web: {
          containerPort: 3000
        }
      }
    }
    runtimes: {
      kubernetes: {
        pod: {
          containers: [
            {
              name: 'log-collector'
              image: 'ghcr.io/radius-project/fluent-bit:2.1.8'
            }
          ]
          hostNetwork: true
        }
      }
    }
  }
}

Remember to save your app.bicep file after you’ve made the above changes.

Step 4: Redeploy your app with the PodSpec added

1. Deploy and run your app again:

   rad run ./app.bicep -a demo
   

   Once the deployment completes successfully, you should see the same deployment completion confirmation message as before, but this time with some system logs from log-collector streaming to your console output:

   Starting log stream...
   
   + demo-547d7dc77f-nmqpk › log-collector
   + demo-547d7dc77f-nmqpk › demo
   demo-547d7dc77f-nmqpk log-collector Fluent Bit v2.1.8
   demo-547d7dc77f-nmqpk log-collector * Copyright (C) 2015-2022 The Fluent Bit Authors
   demo-547d7dc77f-nmqpk log-collector * Fluent Bit is a CNCF sub-project under the umbrella of Fluentd
   demo-547d7dc77f-nmqpk log-collector * https://fluentbit.io
   demo-547d7dc77f-nmqpk log-collector
   

   Access the application by opening http://localhost:3000 in a browser, where you should see the demo app again unchanged from before:

   

   
   When you’re ready to move on to the next step, use CTRL + C to exit the command.

2. Run the command below, which will list the pods in your Kubernetes cluster, using the -o flag to specify the relevant information to output:

   kubectl get pods -A -l app.kubernetes.io/name=demo -o custom-columns=POD:.metadata.name,STATUS:.status.phase,CONTAINER_NAMES:spec.containers[:].name,CONTAINER_IMAGES:spec.containers[:].image
   

   You should now see in the output the original demo app container as before, but also an additional log-collector container that is running in your pod, similar to the following:

   POD                     STATUS    CONTAINER_NAMES      CONTAINER_IMAGES
   demo-547d7dc77f-nmqpk   Running   log-collector,demo   ghcr.io/radius-project/fluent-bit:2.1.8,radius.azurecr.io/tutorial/webapp:latest
   

   Note that you might see old pods with a state of Terminating in the output - this is normal and you should see them disappear once the redeployment completes cleaning up the old resources.

   The log-collector container was deployed using the PodSpec definition you added to your app.bicep file in the runtimes property you added, and is now running alongside your original demo app container.

Cleanup

Run the following command to delete your app and container:

rad app delete demo

Further reading

• Kubernetes in Radius containers
• PodSpec in Radius containers