Open source · 2024
Anchor CLI
Open source · MIT · Author and maintainer
A command-line tool that stands up a production-shaped EKS environment — VPC, node groups, IRSA, ingress, observability and a deploy pipeline — from one config file.
1,240
GitHub stars
40 min
from empty account to running ingress
23
outside contributors
The problem
Every migration I ran repeated the same four days of Terraform archaeology before anything could be deployed. The defaults teams reach for on day one are the ones they regret in month six.
What I did
- Encoded the opinionated defaults: private subnets, IRSA over node roles, PDBs, resource requests, and an ingress with sane timeouts.
- Generated Terraform rather than wrapping it, so teams keep an auditable plan they own.
- Shipped a drift check that flags where a cluster has diverged from the config it was created from.
- Documented every default with the incident that justifies it.
Architecture
Java-based CLI distributed as a native image, generating Terraform modules and Helm values; optional GitHub Actions workflow templates for plan-and-apply with environment gates.
JavaTerraformHelmEKSGitHub Actions
Request path
Similar problem on your side?