The problem
A personal reference for how cloud infrastructure should be built from the ground up — reproducible, declarative, and observable by default, rather than assembled ad hoc project by project.
Infrastructure as code
Reproducible cloud infrastructure on AWS — VPC, subnets, Security Groups, IAM — provisioned through modular Terraform configurations.
Orchestration
Containerized services deployed onto a Kubernetes cluster via custom Helm charts, with resource limits, Secrets, and Horizontal Pod Autoscalers defined explicitly rather than left to defaults.
GitOps
A declarative GitOps pipeline with ArgoCD syncing cluster state against Git repositories, enabling zero-downtime rolling updates.
Monitoring
End-to-end monitoring with Prometheus and Grafana, with proactive alerts for Pod CrashLoopBackOff states and elevated HTTP 5xx rates.
What shipped
- Provisioned reproducible cloud infrastructure on AWS (VPC, subnets, Security Groups, IAM) using modular Terraform configurations.
- Deployed containerized services onto a Kubernetes cluster via custom Helm charts, defining resource limits, Secrets, and Horizontal Pod Autoscalers (HPA).
- Configured a declarative GitOps pipeline with ArgoCD, syncing cluster state against Git repositories to enable zero-downtime rolling updates.
- Established end-to-end monitoring using Prometheus and Grafana, setting up proactive alerts for Pod CrashLoopBackOff states and elevated HTTP 5xx rates.