F5 Distributed Cloud > F5 Distributed Cloud: Container as a Service > Module 4: Monitoring Operating System Metrics with MQTT and Grafana Source | Edit on

Lab 2 - Deploy Containers on vK8s and Build Grafana Dashboard¶

Exercise 1 - Setup Jumphost to connect to the vk8 cluster

Environment Setup To complete this lab section, we’ll need to complete the following steps:

Ensure that the NAMESPACE environment variable is set.
Review the Load Balancer and Origin Server Configuration.
Deploy Grafana using docker compose which will be proconfigured to match your namespace name for each of the 3 regions.

Returning back to the Lab Components view, click the jumphost and then click the Access button. From the access list, select Web Shell.

From Web Shell, run the following commands:

to view the current kubeconfig of the jumphost.

kubectl config view
   ## output shows kubeconfig is not configured for any cluster.
apiVersion: v1
clusters: null
contexts: null
current-context: ""
kind: Config
preferences: {}
users: null

to verify that your file is uploaded.

ls /srv/filebrowser/
 ### Your kubeconfig YAML file you uploaded in tne previous Lab should be listed here:
 ### ves_kind-python_asmith-vk8.yaml

In this example, the NAMESPACE is kind-python and the Virtual K8s is asmith-vk8.

From Web Shell, modify and run the following command to set the NAMESPACE environment variable:

Warning

This step is key to ensure the subsequent commands are configured correctly with your namespace.

Each time you launch a new Web Shell, the namespace environment variables will need to be set.
```
### Replace <namespace> your own namespace value
export NAMESPACE=<namespace>
```
Next, we’ll configure your environment to access the vK8s cluster using the kubeconfig file we uploaded. Choose one of the following methods:
Option 1: Use the ‘cat’ command to configure the kubernetes ‘config’ file.
### Assuming you only have one kubeconfig file in the /srv/filebrowser run: cat /srv/filebrowser/* > ~/.kube/config
Option 2: Use the ‘export’ to set the KUBECONFIG environment variable.
### Otherwise, modify and run: export KUBECONFIG=/srv/filebrowser/<the uploaded file>
Now, let’s view the kubeconfig of the jumphost, again.
```
kubectl config view
```
The output should look like this, showing your vk8s cluster.

Exercise 2 - Deploy Containers on vK8s and Add Origin Pool and Load Balancer

Deploy Containers on vK8s

Now we can deploy the containers into the vK8s cluster. Do this by using the kubectl command to apply the manifest files in the vk8s directory.
```
cd ~/caaslab
kubectl apply -f vk8s/
```
Return to the Distributed Cloud console and in the Distributed Apps workspace select Virtual K8s under Applications.

Click on your vk8 cluster to view the details.
Review all the tabs on your Virtual K8s; Workloads, Deployments, … Pods. To see any hidden tabs on the list, click the button with right pointing chevron.

Questions:

Which tabs show that configurations have been added by the previous command?

Why isn’t there a Workload configured for these Pods?

Exercise 3 - Review the Origin Pool and the TCP Load Balancer Configuration

On the Distributed Cloud console and in the Multi-Cloud App Connect workspace, under Manage, hover over Load Balancers, then click Origin Pools.
Under the Actions menu, for the row adjective-animal-origin click the … and select Manage Configuration.

Note that this origin pool is referencing a K8s service called mosquitto.adjective-animal, and is associated with the Virtual Site appworld2025-k8s-vsite.

We’ve also configured the Origin Pool to use the Endpoint Selection as Local Endpoints Only. This means that the Origin Pool will only use the local endpoints in the region where the Origin Pool is configured and will not cross regions. This is useful when you want to ensure that traffic stays local to the region.
Next, let’s review the TCP Load Balancer which points to the Origin Pool we just saw.

In the Distributed Cloud console and in the Multi-Cloud App Connect workspace, under Manage, hover over Load Balancers, then click TCP Load Balancers.
Again, under the Actions menu, for the row adjective-animal-lb for the TCP Load Balancer, click the … and select Manage Configuration.

The TCP Load Balancer is configured to use the Origin Pool we just reviewed.

Your load balancers is configured to listen 3 different names:
- keen-duck.useast.lab-app.f5demos.com
- keen-duck.europe.lab-app.f5demos.com
- keen-duck.uswest.lab-app.f5demos.com
The LB is also configured to listen on port 8883 and is using SNI.

From the following image, in the Custom Advertise VIP Configuration, we are advertiseing this VIP to the Internet using the virtual site appworld2025-k8s-vsite. This will advertise our MQTT service on each of our regions to the Internet.

Exercise 4 - Deploy Grafana

In this section, we will deploy Grafana using docker compose. The Grafana dashboard will be preconfigured to match your namespace name for each of the 3 regions.

Our docker compose configuration will deploy Grafana with 3 datasources, one for each region. It will also deploy a Dashboard that will show the system stats for each region using the 3 datasources.

Return to the Web Shell and run the docker command to see what containers are already running.
```
docker ps -a
```
We see that docker is already running with some containers, but Grafana is yet to be deployed.
Let’s view the docker-grafana.yaml file to see what will be deployed.
```
cd ~/caaslab/docker-grafana
cat docker-compose.yaml
```
Notice that the environment NAMESPACE variable we set earlier will be used as a docker variable when creating the Grafana container.
To bring up Grafana, from Web Shell run the following commands:
```
cd ~/caaslab/docker-grafana
docker compose up -d
```

Continue to Lab3 to access Grafana, publish data to the MQTT broker and view the dashboard.

Optional: If you have time, review the folders under the “provisioning” directory to find the configuration files for the datasources and the dashboard.

Previous Next