Develop Edge Device Metrics Export Tool for Kubernetes (Simulated)

This guide provides the steps and commands necessary to set up a simulated edge device metrics export tool that integrates with a Kubernetes-native monitoring system (Prometheus ecosystem).

Tech Stack: Basics of YAML, Command-line, Simple Python/Go (Simulated)

---

Core Components

The standard approach uses the Prometheus ecosystem, with the tool acting as a custom metrics exporter.

• Edge Device Simulator (Python/Go): A script that generates simulated metrics (e.g., cpu_usage, memory_allocated).
• Kubernetes Manifests (YAML): Configuration files to deploy the simulator and necessary services in Kubernetes.
• Command-line Interface (CLI): kubectl and possibly curl for deployment and verification.

Note: Prometheus is a system that collects, stores, and queries time-series metrics.

Key Design Constraints:

• ✅ Only 1 Python file (The metrics server: exporter.py)
• ✅ Only 1 Kubernetes YAML file (The deployment: edge-exporter.yaml)
• ✅ No Prometheus client library (Simple text output)
• ✅ No DaemonSet, only a single Pod (Simplest deployment)
• ✅ Works with curl, kubectl, and Docker

---

1. Prerequisites & Minikube Setup

Minikube starts a single-node Kubernetes cluster for local development.

1.1 Minikube Installation

brew install minikube

1.2 Start Minikube & Check Status

minikube start
kubectl get nodes

You should see the node status:

NAME          STATUS   ROLES           AGE   VERSION
minikube      Ready    control-plane   53s   v1.34.0

1.3 Configure Local Docker Environment (Crucial Step)

Make Minikube use your local Docker daemon so the image you build is accessible inside the cluster.

eval $(minikube docker-env)

---

2. Build & Deploy the Exporter

2.1 Build the Docker Image

Build the container image using the Dockerfile in the current directory.

docker build -t edge-metrics .

2.2 Create the Kubernetes YAML File

You must have the edge-exporter.yaml file ready. This manifest defines:

• A Deployment (to run the Pod)
• A Service (to expose the metrics port 8080 internally, annotated for Prometheus scraping)

2.3 Deploy to Kubernetes

Run the following command to deploy the resources defined in the YAML file:

kubectl apply -f edge-exporter.yaml

2.4 Check Pod Status

Check the status of the Pod deployed by your Deployment.

kubectl get pods

Simple Definition: A Pod is the smallest deployable unit in Kubernetes. It's a box that usually contains one container (your Python metrics exporter).

You should see output similar to this:

NAME                                     READY   STATUS    RESTARTS   AGE
edge-exporter-deployment-xxxxx           1/1     Running   0          5s

2.5 Forward Port to Check Metrics (Verification)

Find the exact Pod name and forward its port 8080 to your local machine to ensure the metrics server is live and responding. Prometheus-style metrics must change every time you refresh.

Find the Pod name:

kubectl get pods

Forward the port: (Open a dedicated terminal window)

kubectl port-forward <pod-name> 8080:8080

Test the output: (Open a new terminal window)

curl http://localhost:8080/metrics

---

3. Connect to Prometheus (Monitoring)

STEP 3.1 — Access Prometheus UI inside Minikube

Run this command to automatically open the Prometheus UI in your browser:

minikube service prometheus-server

Note: Keep this terminal window open — it handles the connection/proxy.

STEP 3.2 — Verify your exporter is being scraped

In the Prometheus UI, go to: Status → Targets.

Look for your edge-exporter-service pod in the table.

It should show State = UP.

If it shows DOWN, check the pod logs:

kubectl logs <pod-name>

STEP 3.3 — Query your metrics in Prometheus

Go to the Graph tab in the Prometheus UI.

In the expression box, type the name of one of your custom metrics:

cpu_usage

Click Execute.

Switch to Graph view: you should see metrics updating.

Repeat with your other metric:

memory_allocated

---

Project Structure

project/
├── exporter.py              # Python metrics exporter script
├── edge-exporter.yaml       # Kubernetes deployment manifest
├── Dockerfile               # Container image definition
└── README.md                # This file

---

Troubleshooting

Common Issues

Issue: Pod shows ImagePullBackOff error

• Solution: Ensure you ran eval $(minikube docker-env) before building the Docker image

Issue: Metrics endpoint returns connection refused

• Solution: Verify the pod is running with kubectl get pods and check logs with kubectl logs <pod-name>

Issue: Prometheus shows target as DOWN

• Solution: Check that the Service is properly configured and the pod is exposing metrics on port 8080

Issue: minikube service prometheus-server fails

• Solution: Ensure Prometheus is installed in your cluster. You may need to install it using Helm or kubectl apply

---

Cleanup

To remove all deployed resources:

kubectl delete -f edge-exporter.yaml

To stop Minikube:

minikube stop

To delete the Minikube cluster:

minikube delete

---

Next Steps

• Add more realistic metrics (temperature, network throughput, disk I/O)
• Implement custom labels for multi-device scenarios
• Create alerting rules in Prometheus
• Add Grafana dashboards for visualization
• Scale to multiple edge device simulators

---

Additional Resources

• Doc Link : https://docs.google.com/document/d/1X-oaQYzDJqMBkOdPitBkrnm5JZRhxBEBKUNdarT7dxU/edit?tab=t.0
• Prometheus Documentation
• Kubernetes Documentation
• Minikube Documentation
• Writing Exporters for Prometheus

---

Images

metrics-endpoint

prometheus-cpu-usage-graph

prometheus-memory-graph