A small end-to-end IoT data pipeline demonstrating containerized microservices on Kubernetes with FastAPI ingestion, MongoDB storage, and a background telemetry processor.

This project demonstrates a containerized IoT data pipeline deployed on Kubernetes, including a FastAPI ingestion service, MongoDB storage, a background telemetry processor, and a Python-based device simulator.

• Python (FastAPI, PyMongo)
• MongoDB
• Docker
• Kubernetes (kind)
• Terraform (cloud IaC example)

flowchart LR

subgraph Local_Environment
    SIM[IoT Device Simulator]
end

subgraph Kubernetes_Cluster
    API[FastAPI Ingestion API]
    PROC[Telemetry Processor Service]
    DB[(MongoDB Database)]
end

SIM -->|HTTP Telemetry| API
API -->|Store Raw Telemetry| DB
PROC -->|Read Raw Telemetry| DB
PROC -->|Write Processed Telemetry| DB

project-6-iot-k8s-platform/
  README.md
  terraform/
  k8s/
  services/
    api/
    processor/
  simulator/

kind create cluster --name iot-local

kubectl apply -f k8s/namespace.yaml

kubectl apply -f k8s/mongodb.yaml

Verify:

kubectl -n iot-platform get pods

docker build -t iot-api:0.1 ./services/api
kind load docker-image iot-api:0.1 --name iot-local
kubectl apply -f k8s/api.yaml

docker build -t iot-processor:0.1 ./services/processor
kind load docker-image iot-processor:0.1 --name iot-local
kubectl apply -f k8s/processor.yaml

kubectl -n iot-platform port-forward svc/iot-api 8000:8000

In another terminal:

python3 simulator/send_telemetry.py --devices 3 --interval 1 --rounds 5

1. Simulator sends telemetry via HTTP
2. FastAPI ingestion service stores raw telemetry in MongoDB
3. Processor service reads new telemetry
4. Processor classifies data (HOT, COLD, HUMID, NORMAL)
5. Processed data stored in a separate MongoDB collection

The terraform/ folder contains an example Infrastructure as Code configuration for provisioning cloud infrastructure (e.g., VPC + networking skeleton).

This demonstrates:

• Infrastructure as Code practices
• Cloud-native thinking
• Separation of local development and cloud deployment patterns

• Microservices architecture on Kubernetes
• Docker image build & cluster loading
• Kubernetes Deployments, Services, Secrets, PVC
• NoSQL data storage with MongoDB
• Background data processing service
• End-to-end telemetry simulation
• DevOps workflow using kind
• Infrastructure as Code awareness (Terraform)

• Add Kafka for real message streaming
• Add Prometheus + Grafana monitoring
• Deploy to managed Kubernetes (EKS/GKE/AKS)
• Add CI/CD pipeline (GitHub Actions)
• Add Horizontal Pod Autoscaler

Pattarin Thunyapar
Master’s in Data Analytics
Focused on Data Engineering, Cloud, and AI Systems