Deploy Kubernetes Cluster on AWS
Budget: €12 – €18 EUR
I need a production-ready Amazon EKS cluster spun up in my AWS account and documented so my team can take over day-to-day operations confidently. The goal is a secure, highly available Kubernetes environment that follows AWS best practices and leaves no “snowflake” resources behind.
Scope
– Provision the entire stack (VPC, subnets, IAM roles, node groups, autoscaling, logging) through infrastructure-as-code. I’m comfortable with Terraform, but I’m fine if you prefer eksctl as long as everything is reproducible.
– Configure cluster access controls (RBAC, IAM roles for service accounts) and network policies so that workloads are isolated and the principle of least privilege is respected.
– Enable observability: CloudWatch metrics and logs wired into the cluster; open to Prometheus if you think it adds value.
– Set up a basic deployment pipeline that lets me push a Docker image and see it roll out to the cluster automatically. GitHub Actions is our current CI/CD tool, so please wire the workflow to EKS.
– Hand over concise documentation: repo README, diagram of the architecture, and a short run-book for scaling or disaster recovery.
Acceptance criteria
1. All resources can be recreated from the IaC in one command.
2. kubectl context works from my workstation and from a GitHub Actions job.
3. A sample “hello-world” service deploys successfully and is reachable through an ALB Ingress.
4. CloudWatch (or Prometheus) shows node and pod metrics within five minutes of deployment.
5. Final hand-off call confirms access, monitoring, and rollback procedures.
I’ll grant temporary, least-privilege IAM access and expect the project to stay within that scope. If you’ve built EKS clusters before and can demonstrate previous setups or repos, that will help me gauge fit quickly.
Scope
– Provision the entire stack (VPC, subnets, IAM roles, node groups, autoscaling, logging) through infrastructure-as-code. I’m comfortable with Terraform, but I’m fine if you prefer eksctl as long as everything is reproducible.
– Configure cluster access controls (RBAC, IAM roles for service accounts) and network policies so that workloads are isolated and the principle of least privilege is respected.
– Enable observability: CloudWatch metrics and logs wired into the cluster; open to Prometheus if you think it adds value.
– Set up a basic deployment pipeline that lets me push a Docker image and see it roll out to the cluster automatically. GitHub Actions is our current CI/CD tool, so please wire the workflow to EKS.
– Hand over concise documentation: repo README, diagram of the architecture, and a short run-book for scaling or disaster recovery.
Acceptance criteria
1. All resources can be recreated from the IaC in one command.
2. kubectl context works from my workstation and from a GitHub Actions job.
3. A sample “hello-world” service deploys successfully and is reachable through an ALB Ingress.
4. CloudWatch (or Prometheus) shows node and pod metrics within five minutes of deployment.
5. Final hand-off call confirms access, monitoring, and rollback procedures.
I’ll grant temporary, least-privilege IAM access and expect the project to stay within that scope. If you’ve built EKS clusters before and can demonstrate previous setups or repos, that will help me gauge fit quickly.
Related categories:
Linux
Cloud Computing
Amazon Web Services
Network Administration
Docker
Kubernetes
Terraform
CI/CD