Plane AWS Deployment Kit — Automated fork-and-deploy with Terraform, Helm & GitHub Actions

[Semih702]

Hi everyone,

I'm a computer science student, and I built an open-source AWS deployment kit for Plane Community Edition. It started as a cloud computing course project and grew into a complete, fork-friendly solution after I noticed there was no end-to-end Terraform + Helm + GitHub Actions workflow in the existing self-hosting options.

What it does

Fork the repo, add one GitHub secret (your AWS OIDC role ARN), and run two workflows. No long-lived AWS access keys needed — the default path uses GitHub OIDC and IRSA throughout.

You can deploy into a new dedicated EKS cluster, or set CREATE_EKS_CLUSTER=false to install Plane into an existing cluster without duplicating infrastructure. Every dependency (PostgreSQL, Redis, RabbitMQ, object storage) can be independently switched between AWS-managed, self-managed in-cluster, or externally hosted.

The default path provisions:

• VPC, EKS cluster, managed node group
• RDS PostgreSQL, ElastiCache Redis, Amazon MQ RabbitMQ
• S3 doc-store with IRSA (no static credentials)
• AWS Load Balancer Controller, Cluster Autoscaler, metrics-server
• Plane CE via a wrapper Helm chart (built on the official plane-ce chart)
• Generated secrets stored in AWS Secrets Manager

Configurable service modes

Each dependency can be AWS-managed, in-cluster, or external:

Service	Modes
PostgreSQL	aws-managed · in-cluster · external
Redis	aws-managed · in-cluster · external
RabbitMQ	aws-managed · in-cluster · external
Object Store	s3-managed · minio-in-cluster · external-s3

Organizations with an existing EKS cluster can set CREATE_EKS_CLUSTER=false and deploy Plane into a namespace without creating a second cluster.

Relation to existing AWS tooling

I am aware of the commercial-deployments repository, which targets commercial Plane deployments using Terraform and Kustomize. This kit takes a complementary approach rather than a competing one — it focuses on the Community Edition, uses Helm (building on the official plane-ce chart from helm-charts), and wraps everything in a GitHub Actions CI/CD pipeline so that a fork is immediately deployable. It could serve as an alternative entry point for teams that prefer a Helm-based, fully automated workflow for their AWS self-hosting setup.

Links

• Repository: https://github.com/Semih702/plane-cloud-platform
• Detailed proposal: https://github.com/Semih702/plane-cloud-platform/blob/main/docs/proposal-to-plane.md

Future ideas

If there is interest in taking this further, I have a few ideas I would like to explore:

• Automated version tracking with Renovate: keep EKS, Helm chart, and provider versions up to date through automated pull requests
• Monitoring stack: optional Prometheus + Grafana integration for Plane workloads
• Cost estimation: surface estimated AWS costs before apply based on selected service modes
• Multi-cloud support: extend the kit to support other cloud providers (GCP, Azure) alongside AWS, so teams can deploy Plane on their preferred platform

These are just starting points; I would be happy to align on a roadmap together if the Plane team finds this direction useful.

I would love feedback from the Plane team and community. If this fits the self-hosting direction, I am happy to maintain it long-term and adapt it to Plane's standards, whether as a docs reference, a community repo under the Plane org, or any other form.