Zadara Cloud + EKS + DataDog: Transforming the Pacman Game into a Professional Kubernetes Platform

Written by Osmar R Leão

SA @Zadara

---

Continuing the articles about Zadara Cloud and Kubernetes, this final(ish) article will demonstrate a professional observability tool for the professional EKS environment. I choose to work with DataDog but it could be any professional tool that integrates with Kubernetes.

The goal is to observe all the Kubernetes metrics to gather information and make decisions based on real data. Datadog is a handy and powerful tool for this. I look forward to an official partnership in the future.

After the integration, it will be possible to view metrics about your cluster(s) on the Zadara Cloud:

Cluster overview on DataDog

On the Explorer tab is possible to navigate through the Kubernetes resources:

Kubernetes pod overview

NOTE: observe the AWS Load Balancer Controller pods and the EBS CSI pods to allow AWS-compatible deployments to run on the Zadara Cloud (remember it’s an EKS Distribution).

Configuration Process

The installation and configuration process from the DataDog website is complete, but, to work with EKS on top of zCompute, is needed some adjustments:

Agent Installation Instructions from DataDog website

The default YAML file from the agent installation wizard is:

apiVersion: "datadoghq.com/v2alpha1"kind: "DatadogAgent"metadata:  name: "datadog"spec:  global:    clusterName: "zadara-demo"    credentials:      apiSecret:        secretName: "datadog-secret"        keyName: "api-key"

And, to work with EKS on the zCloud must be like:

apiVersion: "datadoghq.com/v2alpha1"kind: "DatadogAgent"metadata:  name: "datadog"spec:  global:    clusterName: "zadara-demo"    credentials:      apiSecret:        secretName: "datadog-secret"        keyName: "api-key"    kubelet:      tlsVerify: false  override:    nodeAgent:      env:        - name: DD_HOSTNAME          valueFrom:            fieldRef:              fieldPath: spec.nodeName

These modifications are important to get things working smoothly.

The agent pods will start and a few moments later should be running:

$ kubectl get podsNAME                                     READY   STATUS    RESTARTS   AGEdatadog-agent-44brf                      3/3     Running   0          47hdatadog-agent-cg6c5                      3/3     Running   0          47hdatadog-agent-s4d7q                      3/3     Running   0          47hdatadog-cluster-agent-648764f496-ztm28   1/1     Running   0          47hdatadog-operator-866c9f4cf6-bg5s4        1/1     Running   0          2d5h

Now, all the work is using the DataDog web interface.

Navigating the Interface

Pods:

list of running pods

Cluster:

list of clusters

Namespaces:

namespaces of the cluster

Nodes:

nodes of the Kubernetes

Deployments:

deployments done

These deployments included the Pacman game, one app using an SSL demo (showed in my first article about EKS on zCloud), the MongoDB for the Pacman game, the DataDog itself, and the Kubernetes and EKS deployments.

Services:

services running

The MongoDB and Pacman Load Balancer from my second article.

Ingresses:

application load balancer

The ingress is integrated with zCloud, and using SSL certificate uploaded to the zCompute console. The class is ALB, so it is an Application Load Balancer (the creation was demonstrated in my first article).

Storage (Persistent Volume Claims and Persistent Volumes):

active PVCs

volumes created

Details and Dashboards

It is possible to view all the details about each item listed above by clicking on it. The detail of the Mongo PV, for example:

all the details of mongodb volume created

On the tab “Resource Utilization” it is possible to view CPU and Memory consumption by each item:

memory consumption of the pods

CPU consumption of the pods

Dashboards:

The DataDog allows the creation of dashboards to monitor all aspects of the infrastructure:

list of pre-formatted dashboards

Here are some examples:

It is possible to create a customized dashboard mixing information, below is an example of Pacman monitoring:

Pacman monitoring 01

Pacman monitoring 02

Pacman monitoring 03

Conclusion

The Zadara Compute Infrastructure, combined with EKS distribution and monitored by DataDog, creates a robust, professional Kubernetes platform.

Zadara Cloud’s key strength lies in its ability to combine powerful solutions with a tailor-made Federated Edge Cloud, delivering high performance and cost efficiency while staying close to customers.

---