I created a HorizontalPodAutoscaler in helm charts to automatically scale workload per the average CPU and Memory utilization of target pods to match demand
I recommend a Serverless Posgress Database instance to accomodate unpredictabled production workload and to scale with zero downtime in response to increased resource demand
The application is going to be hosted on a three-tier layer architecture and multi-region setup
Tier 1: Client App Tier 2: Application Server Tier 3: Database Server
The infrastructure will be deployed with Terraform (Iac) to provision required infrastructure resources to automate, avoid misconfiguration, and reproduce system configuration.
- Virtual Network
- Virtual Machines
- Database Instance
- Loadbalancer
- API Gateway
3 nodes 8vCPU, 16 GB RAM will be provisioned, and the Kubernetes cluster will be bootstrapped with KubeAdm with Ubuntu 24.04 as the base OS and Kubernetes version v1.30.3.
- 1 Control-Plane Node
- 2 Worker Nodes
Argo CD is a declarative, GitOps continuous delivery tool for Kubernetes
GitHub Actions is a continuous integration and continuous delivery (CI/CD) platform to automate the build, test, and deployment pipeline
-
GitOps framework to manage helm chart repository and automate deployment to Kubernetes cluster
-
A CI/CD Pipeline with Docker to build application code and push container image to docker hub registry
Tools: Terraform, Kubernetes, Helm, Docker, Github Actions, ArgoCD